Underwater technology 35 1 by SUT

Vol. 35 32 No. No. 132 2018 2014 Vol.

UNDERWATER TECHNOLOGY

A Personal View... Tides of change and opportunity

Steve Hall

Errors in pressure to depth approximations and how to avoid them

Jose M Puig

Design and implementation of a remotely operated vehicle testbed

De-noising algorithm for SNR improvement of underwater acoustic signals using CWT based on Fourier Transform

ISSN 1756 0543

Kishore Kumar PC, Sathish Kumar P, Jerritta S and Rajendran V

Book Review A Farewell to Ice

Zhigang Deng, Fang Yuan, Daqi Zhu and Feng Xue

www.sut.org

0A-SUT-35(1)-OFC.indd 1

3/27/18 1:57 PM

UNDERWATER TECHNOLOGY Editor Dr MDJ Sayer Scottish Association for Marine Science Assistant Editor E Azzopardi SUT Editorial Board Chairman Dr MDJ Sayer Scottish Association for Marine Science Gavin Anthony, GAVINS Ltd Dr MA Atamanand, National Institute of Ocean Technology, India LJ Ayling, Maris International Ltd Commander Nicholas Rodgers FRMetS RN (Rtd) Prof Ying Chen, Zhejiang University Jonathan Colby, Verdant Power Neil Douglas, Viper Innovations Ltd, Prof Fathi H. Ghorbel, Rice University G Griffi ths MBE, Autonomous Analytics Prof C Kuo FRSE, Emeritus Strathclyde University Dr WD Loth, WD Loth & Co Ltd Craig McLean, National Ocean and Atmospheric Administration Dr S Merry, Focus Offshore Ltd Prof Zenon Medina-Cetina, Texas A&M University Prof António M. Pascoal, Institute for Systems and Robotics, Lisbon Dr Alexander Phillips, National Oceanography Centre, Southampton Prof WG Price FRS FEng, Emeritus Southampton University Dr R Rayner, Sonardyne International Ltd Roland Rogers CSCi, CMarS, FIMarEST, FSUT Dr Ron Lewis, Memorial University of Newfoundland Prof R Sutton, Emeritus Plymouth University Dr R Venkatesan, National Institute of Ocean Technology, India Prof Zoran Vukic´, University of Zagreb Prof P Wadhams, University of Cambridge Cover Image (top): zoonar.com/syrist Cover Image (bottom): Steve Crowther Cover design: Quarto Design/ kate@quartodesign.com

Society for Underwater Technology Underwater Technology is the peer-reviewed international journal of the Society for Underwater Technology (SUT). SUT is a multidisciplinary learned society that brings together individuals and organisations with a common interest in underwater technology, ocean science and offshore engineering. It was founded in 1966 and has members in more than 40 countries worldwide, incIuding engineers, scientists, other professionals and students working in these areas. The Society has branches in Aberdeen, London and South of England, and Newcastle in the UK, Perth and Melbourne in Australia, Rio de Janeiro in Brazil, Beijing in China, Kuala Lumpur in Malaysia, Bergen in Norway and Houston in the USA. SUT provides its members with a forum for communication through technical publications, events, branches and specialist interest groups. It also provides registration of specialist subsea engineers, student sponsorship through an Educational Support Fund and careers information. For further information please visit www.sut.org or contact: Society for Underwater Technology 1 Fetter Lane EC4A 1BR London UK e info@sut.org t +44 (0)20 3440 5535 f +44 (0)20 3440 5980

Scope and submissions The objectives of Underwater Technology are to inform and acquaint members of the Society for Underwater Technology with current views and new developments in the broad areas of underwater technology, ocean science and offshore engineering. SUT’s interests and the scope of Underwater Technology are interdisciplinary, covering technological aspects and applications of topics including: diving technology and physiology, environmental forces, geology/geotechnics, marine pollution, marine renewable energies, marine resources, oceanography, salvage and decommissioning, subsea systems, underwater robotics, underwater science and underwater vehicle technologies. Underwater Technology carries personal views, technical papers, technical briefings and book reviews. We invite papers and articles covering all aspects of underwater technology. Original papers on new technology, its development and applications, or covering new applications for existing technology, are particularly welcome. All papers submitted for publication are peer reviewed through the Editorial Advisory Board. Submissions should adhere to the journal’s style and layout – please see the Guidelines for Authors available at www.sut.org.uk/journal/default.htm or email elaine.azzopardi@sut.org for further information. While the journal is not ISI rated, SUT will not be charging authors for submissions.

in more than 40 countries worldwide, including over 190 Corporate Members of the Society.

Disclaimer and copyright The Society does not accept responsibility for the technical accuracy of any items published in Underwater Technology or for the opinions expressed in such items. The copyright of any paper published in the journal is retained by the author(s) unless otherwise stated. All authors are supplied with a PDF version of their papers once published. Authors are encouraged to make the PDF version of their papers free to download from their own websites.

Open Access Underwater Technology is available as Open Access. PDF versions of all published papers from Underwater Technology may be accessed via ingentaconnect at www. ingentaconnect.com/content/sut/unwt. All issues from Volume 20 (1995) onwards are available as Open Access. The Society for Underwater Technology also encourages Underwater Technology authors to make their papers available online on their personal and/or institutional websites for Open Access. Through this arrangement, the Society supports the Open Access policy not only in the UK (the Research Councils UK (RCUK) policy) but also the drive towards Open Access in other countries.

Abstracting and indexing Underwater Technology is included in Emerging Sources Citation Index. Additional abstracting and indexing services include American Academy of Underwater Sciences (AAUS) E-Slate; Aquatic Sciences and Fisheries Abstracts (Biological Sciences and Living Resources; Ocean Technology, Policy and Non-Living Resources; and Aquatic Pollution and Environmental Policy); Compendex; EBSCO Discovery Service; Fluidex; Geobase; Marine Technology Abstracts; Oceanic Abstracts; Scopus; and WorldCat Discovery Services.

Subscription Subscription to the print version of Underwater Technology is available to non-members of the Society at the following rates per volume (single issue rates in brackets). Prices are given in GBP. Accepted methods of payment are cheque or credit card (MasterCard and Visa). Foreign cheques must be in GBP and drawn on a British bank otherwise a currency conversion surcharge is incurred. UK subscription Overseas subscription

£102.00 (£25.50 per issue) £108.00 (£27.00 per issue)

Underwater Technology is also available in electronic format via ingentaconnect as Open Access. To subscribe to the print version of the journal or for more information please email Elaine Azzopardi at elaine.azzopardi@sut.org

Publication and circulation Underwater Technology is published in March, July and November, in four issues per volume. The journal has a circulation of 2,400 copies to SUT members and subscribers

0B-SUT-35(1)-IFC.indd 1

Advertising To book an advert or for more information please contact Elaine Azzopardi at elaine.azzopardi@sut.org

3/27/18 1:57 PM

www.sut.org

A Personal View...

doi:10.3723/ut.35.001 Underwater Technology, Vol. 35, No. 1, pp. 1–2, 2018

Tides of change and opportunity It’s coming up to a year since I took up the role of Chief Executive Officer of the Society for Underwater Technology, and it has been a fascinating experience learning more about the breadth and capability of our membership. Although I’ve been a Member and Fellow of SUT for many years, my contacts before joining full-time were predominantly from the marine science, technology and education end of our membership spectrum, so I’ve enjoyed getting to know the offshore energy specialists who provide the solid base of our individual and corporate membership. The combined expertise and experience of this broad range of people, companies, institutions and universities puts SUT in a very special place, where we and our sister Societies can together help shape how humankind will interact with the ocean in the future. We find ourselves in a world that is going through one of those great transformation periods that happen every few decades. This time it’s the rise of China and Asia, the global shift of focus away from Europe and North America, the recognition that we must act collectively to solve environmental challenges, and the possibility that we’re returning to an era of protectionism, trade barriers and enhanced national sovereignty over multipartner flexibility. This rapidly changing and hard-to-predict world presents a number of opportunities to: diversify our interests; engage with governments, industry and academics; and build on our strengths as a politically impartial, knowledge-focused Learned

Society that dispenses facts based on evidence, encourages young people to enter a fascinating and necessary sector, and provides a focus for friendly networking, learning and celebration of discovery. In unstable times, our expertise will find itself moving towards new directions, such as investigating how to feed growing human and farm-animal populations, ensuring marine security for our homelands and trade routes, and obtaining an ever-widening range of resources from the ocean floor – not only energy but also minerals and metals that will help us build an electric transport infrastructure. We’ll be playing a part in driving the emergence of new technologies and areas of innovation. First – energy. Offshore hydrocarbons will remain a key fuel for a while yet, making use of new techniques to help access deep or remote resources, to extend the life of existing fields and to access smaller ‘pools’ of oil and gas than were hitherto economically viable. We also have to find efficient ways to decarbonise energy supply in order to reduce greenhouse gas emissions and slow the pace of sea level rise and ocean acidification. We can buy time by replacing coal as a fuel with natural gas, resulting in up to 40 % reduction in carbon output for the same amount of energy generated. The development of carbon capture and storage and other geo-engineering solutions may also enable governments to permit continued oil and gas extraction without breaching carbon reduction targets. But increasingly, the policy driver will be to encourage new energy production

Steve Hall Before joining SUT as CEO Steve Hall spent over 26 years with the Natural Environment Research Council, carrying out a range of duties including tracer chemistry in the wild seas of the Southern Ocean and managing the NERC Autosub Science Missions thematic programme from 1997 to 2002. He spent a decade as a marine policy specialist, becoming Head of the International and Strategic Partnerships Ofﬁce of the National Oceanography Centre, the UK’s and Overseas Territories tsunami warning focal point, and Head of the UK’s delegation to UNESCO’s Intergovernmental Oceanographic Commission, where he was elected Vice-Chair from 2015 to 2017. Steve joined SUT in the 1990s, serving on the Education and Training Committee and later the Policy Advisory Committee, was twice elected to SUT Council serving as Hon Secretary and Chair. He is a Fellow of SUT and IMarEST, a Chartered Marine Scientist and author of children’s books about the ocean and policy brieﬁngs to government on topics from renewable energy and nuclear submarine decommissioning to scallop dredging. When not working for SUT he enjoys cycling and walking with his family in the hills near his home in Wales.

from low or zero-carbon sources such as offshore wind, wave and tidal energy. Even solar power can be produced from floating solar farms, as already deployed in China.

01-SUT299-35(1)-ut66630.indd 1

3/27/18 1:50 PM

Hall. Tides of change and opportunity

However, easily stored and transported fuels such as gas offer such a degree of flexibility that it’s likely we’ll find innovative ways to continue to use gasbased resources, particularly if humanity chooses to travel down the pathway towards a global ‘hydrogen economy’. For hydrogen production, the ocean will be key both as feedstock in the form of seawater for electrolysis to hydrogen and oxygen, and as the medium in which floating wind, nuclear or solar plants to carry out the electrolysis will be deployed. We know there are still huge offshore gas resources, as well as gigatonnes of gas hydrates, and these can also be converted to hydrogen with the carbon removed and pumped into longterm geological storage. It’s not hard to see that there’s plenty of work for SUT members in all of those scenarios! Second – food supply. Feeding the 9 billion plus humans predicted to exist by the middle of this century also poses a substantial challenge, as does feeding our meat and dairy herds. Wild-caught fish can’t be harvested in large enough numbers without a collapse of stock, so aquaculture has to be part of the solution to protein production. It’s reasonable to predict that at least some of that production can be co-located with offshore energy production. Will new companies emerge that derive their income from energy production in its widest sense – fuel for people and animals in the form of food, as well as fuel for transportation, services and heating? Shellfish such as mussels are well-suited to colonise ropes hung from the underside of floating offshore installations. It may be that recent developments in stem-cell based in-vitro meat production will render traditional fishing as obsolete. Could

we see steaming vats of meat cultures quiver ashore and produce tonnes of fish, chicken or other meats without ever having experienced life as part of an actual animal? If that comes to pass, it frees up ocean wildlife stocks to recover to pre-industrial levels. Will lowvolume high-value sea-caught fish become a luxury item at a premium price, perhaps farmed at offshore energy plants? Third – security. A crowded, resource-hungry world is also unfortunately a world where the ability to project power, protect our interests, and demonstrate defensive and offensive capability will become even more important than it is now. Patrol and surveillance will increasingly become the domain of surface and sub-surface marine autonomous systems – a process that is already underway. Perhaps soon, the combat roles will be carried out by robots too, especially as new legal regimes and rules of engagement are developed as a consequence of growing experience and necessity. Although we still don’t know if true artificial intelligence is possible, if it does arrive there will be few limits to what our machine creations – or the machines’ creations – can do. Fourth – minerals, metals and marine genetic resources. Much has been written about the ‘Blue Economy’, and research is underway at universities and institutes across the world to investigate how best to harness new resources of minerals and metals that lie in vast quantities across the ocean floor and mid-ocean ridges without excessive damage to habitat. Some of these resources are the rare-earth metals that will be needed in increasing quantities for the drivetrains of electric cars, trucks and ships. In terms of marine genetic resources, organisms that dwell

in the deep have been found to exhibit new medicinal, nutritional and antibiotic properties. Whilst large-scale harvesting is unlikely (it’s probably easier to synthesise the molecules ashore once discovery has been made), there is work to be done in exploration and developing an understanding of the unique ecosystems of the deep ocean. There, life moves at a slow pace and takes a long time to reach reproductive age, so human disturbance may take them decades to recover from. The UN is currently exploring new ocean governance systems for biological resources located in areas beyond national jurisdiction. SUT will have a role in providing unbiased technical advice to the new agencies that arise from this intergovernmental process, as we already do in the fields of marine technology and knowledge exchange as official observers to the Intergovernmental Oceanographic Commission. Our members will often be the individuals and organisations who pioneer the development of the ‘Blue Economy’, and we’ll need to work hard to ensure that we can develop resources, extract energy and harvest the riches of the ocean without causing unsustainable harm. Our role as educators and supporters of students will be of tremendous value in helping encourage young people to enter the sector, and helping our homelands to deliver the UN Sustainable Development Goals. The Society for Underwater Technology has never been needed more than it is now, to help steer the course through turbulent waters and deliver the kind of world and oceans we want our grandchildren to thrive in. It’s an exciting time and full of possibilities. I’m looking forward to what the future brings.

01-SUT299-35(1)-ut66630.indd 2

3/27/18 1:50 PM

doi:10.3723/ut.35.003 Underwater Technology, Vol. 35, No. 1, pp. 3–12, 2018

Technical Paper

www.sut.org

Errors in pressure to depth approximations and how to avoid them Jose M Puig* Singapore Received 7 December 2017; Accepted 5 February 2018

Abstract Pressure and density based approximations are widely used in the offshore hydrographic industry as a simple and quick way to compute depth from pressure observations. All these approximations are based on the hydrostatic pressure equation, which has no exact solution form for ocean applications; however, it can be solved with numerical integration techniques. The accuracy limitations of pressuredepth approximations have always been historically accepted, because when they were implemented some decades ago, pressure sensors were not that accurate, density proﬁlers were expensive and computational power was limited. Presently, computer and pressure sensor technology have advanced tremendously and water proﬁlers that measure conductivity, temperature and depth (CTD) are now common instruments. To take full advantage of these improved measurement accuracies, the water density should be used in the correct way. This article reviews the different pressure and density based approximations, and compares them to a numerical solution of the hydrostatic equation. Ten CTD casts from around the world are used as test cases. The comparisons suggests that pressure and density based approximations should no longer be used when seeking to compute an absolute depth. They should instead be replaced by the use of the numerical solution to the hydrostatic equation or the complete UNESCO pressureto-depth formula that considers density, both of which offer similar results. For relative depth measurements like subsea spool metrology, the results suggest that the errors of using approximations fall below required measurement tolerances. Key Words: pressure, hydrostatic, depth, UNESCO, hydrographic, survey

1. Introduction High-accuracy absolute-depth measurements are increasingly important in geophysical and geotechnical

* Corresponding author. Email address: jose.puig@sonardyne.com

studies. Some examples are plate tectonics movement and reservoir subsidence monitoring, where centimetre changes in depth are sought (Doornhof et al., 2006). In subsea construction and inspection operations, absolute depth is also required for tracking divers, remotely operated vehicles (ROVs), autonomous underwater vehicles (AUVs) and structures. The required accuracies for tracking can be less rigorous and depend on the type of work undertaken. The International Marine Contractors Association (IMCA) provides information on survey guidance and standards. Relative depth surveys – the subsea equivalent of a height survey – also require accurate depth determination. For example, spool piece metrology requires a determination of connector-to-connector depth differences to design a precisely tailored flexible pipe connector, or a depth loop survey to determine long base line (LBL) acoustic transponder array depths. The expected accuracy is typically in the order of the measurement sensor accuracy (Barret et al., 2017). Measuring accurate and precise absolute depth is a complicated endeavour. The underlying concept of pressure-to-depth conversion is hydrostatic pressure, where a pressure measured at a location in the ocean (x,y,z) at time (t) can be expressed as (LaCasce, 2013): p(x,y,z,t) = p0(z) + p′(x,y,t,z) + pa(x,y,z,t)

(1)

where the x and y represent horizontal coordinates, z is the vertical coordinate with z = 0 at the surface. The hydrostatic pressure (p0) is a function of the pressure exerted by the weight of the stationary water mass above the location (x,y,z), quantified by the height (z) of the water column; p′ is the dynamic component of pressure caused by the oceans constant movement, and captures all of the departures

02-SUT297-35(1)-ut66631.indd 3

3/27/18 1:50 PM

Puig. Errors in pressure to depth approximations and how to avoid them

of pressure from the static state. Finally, pa is the atmospheric pressure as both functions of space and time. Computing the weight of water requires an estimate of gravity (g (x,y,z,t)) and water density (σ(x,y,z,t) – both complicated functions. In an effort to simplify the mathematical problem, the hydrostatic pressure equation ignores the dynamic component and makes simplifying assumptions; for example, it is frequent practice in literature to make pa = 0 at sea level. However, a pressure sensor in the real world will measure the combination of all pressures as per equation 1. In addition, the pressure from the dynamic component of pressure can be small if the pressure measurement is designed correctly. For real-time application, sampling rate and length are used to filter out high frequency ocean oscillations, but there are many more long-term oscillations that cannot be mitigated and can contribute to a realtime pressure measurement. In some regions of the world, loop currents and global weather patterns can change depth significantly, as much as 4 m at 5000 m water depth (Intergovernmental Oceanographic Commission (IOC), 2010). Hydrostatic pressure-to-depth conversion is formally expressed as (Resnick et al., 2001): gdz d =

dp ρ

(2)

where (dz) represents change in depth and is directly related to the change in pressure (dp) owing to the weight of water as described by density rho (ρ) and gravity (g). The density of water is defined as its weight per unit volume and is characteristically a function of the water’s dissolved mineral content or salinity (s)†, its temperature (t) and the pressure (p) exerted by the weight of water column. Sea water density varies greatly in the ocean, in depth, geographically and in time. This paper will use the Equation of State (EoS) to describe it (Foffonov and Millard, 1983; for a good introductory reference to ocean density, see Pickard and Emery, 1990). It is worth mentioning that there is a new formulation known as TEOS-10 for density that is not considered in this paper (refer to the IOC Manual No 56, 2010, and Pawlowicz, 2010, for details). Gravity is a function of height from the centre of the Earth and its overall density mass, which varies with geographic location. In literature, it is considered a function of height (z) from the centre of the Earth and latitude phi (ø) or g = f(z, ∅). Even with these variables accounted for, there is no solution in classical terms to equation 2, and numerical integration methods must be used. † typically measured through water conductivity (c).

The measurements required to compute depth are hydrostatic pressure, a water density profile estimate and the latitude of the location. Pressure is measured with strain gauge or piezoelectric (quartz-based) sensors, and the density profile is estimated from conductivity, temperature and depth (CTD) sensor measurements, although more appropriately it should be called a CTP as its logged raw measurements are conductivity, temperature and pressure. Pressure sensors have evolved and improved over time. The first models were not as accurate as modern sensors and were considerably more expensive. The accuracy of the first strain gauge pressure sensors used in the ocean were typically around 0.1 % of the total pressure scale (Saunders and Foffonov, 1976), compared to modern quartz sensors whose accuracy is 0.01 % or better (Paroscientific, n.d.). Recent progress in understanding pressure sensor drift and techniques to remove it in real time will provide an order of magnitude increase in current pressure sensor accuracy (Paros and Kobayashi, n.d.). In the late 1960s and early 1970s, routine absolute depth measurements became possible because of the increasing development of ocean sensor technology and the increased drive associated with offshore hydrocarbon developments. However, CTDs were still considered exotic instrumentation and offshore computer processing power was minimal. Hence it made sense then to seek simple approximations to equation 2 based on pressure readings alone, which could be handled by simple calculators or tabulations, and whose accuracy was better than the pressure sensor data they used. Technology has advanced greatly since then; modern quartz pressure sensors are relatively inexpensive and very accurate. Computing power is powerful and economical; and CTD sensors are basic fit equipment on every offshore survey package. Despite all this, the simple pressure based depth approximations are still presently used in the offshore industry. The objective of this work is to present a short background of the pressure to depth conversion problem, its origin, history, and the most common approximations used in offshore construction survey operations. A numerical solution to the hydrostatic pressure relationship is presented and used to compare the accuracy of pressure-to-depth conversion approximations. A collection of ten CTP profiles, gathered from various oceans, are used as test cases. The observed differences are discussed and some conclusions are presented.

02-SUT297-35(1)-ut66631.indd 4

3/27/18 1:50 PM

Underwater Technology Vol. 35, No. 1, 2018

This work aims to provide some insight on how complicated it is to measure an absolute depth. It also intends to help create reasonable expectations for depth determinations and, in the longer term, inspire the adoption of an industry standard for depth determination that can perhaps improve the quality and repeatability of current depth measurements practices.

2. Development of pressure depth computations Most approximate solutions to equation 2 are based on the work of Saunders and Foffonov (1976), whose reasoning is briefly as follows: Integration of equation 2 yields: z=∫

p 0

dp gρ

(3)

Integration is from the sea surface where the hydrostatic pressure is assumed to be zero. In practical terms, this is achieved by removing the atmospheric pressure reading from the pressure sensor when at sea level prior to seabed deployment. Assuming a gravity model that is weakly and linearly dependent on pressure: g = g0 + γ′z

(4)

and with the further simplification of replacing pressure for depth (or p for z), it can be implemented in to equation 3. The error introduced to the depth calculation from this approximation is estimated to be 5 cm in 10 000 m water depth (Saunders and Foffonov, 1976). The gravity model above is very simple and ignores gravity anomalies that are estimated to impact changes of up to 0.5 m in 5000 m of water (Saunders and Foffonov, 1976). Substitution of above gravity model in to equation 2 yields: z(m ) =

∫

p 0

αdp

(5)

g + 21 ′ p

where g0 (expressed in m/s2) is the surface gravity dependent on latitude, ø (in degrees) and gamma (γ) is the average gravity gradient with pressure (p) in kg/m2. Alpha (α) represents the specific volume of seawater and is simply the inverse of its density (ρ) in kg/m3). It is a common way of expressing density in oceanographic literature. Equation 5 can be solved using numerical integration using a measured density profile and a pressure measurement. For discussion purposes, this paper will refer to this equation as the hydrostatic depth numerical solution. What follows are approximate solutions to equation 3.

Saunders and Foffonov (1976) presented an approximation of equation 5 based solely on pressure that would remove the need to measure density, and then added a tabulated corrective term for specific ocean basins. They proposed the use of a combined specific volume with two principal contributing terms: ∝ = ∝0 + ∂

(6)

where ∝0 describes an oceanographic concept called the ‘standard ocean’ specific volume, and ∂ describes the deviation from this base state, or formally the specific volume anomaly for a particular ocean basin. Applying this to equation 3: z=

∫

p 0

α0dp

g +

1 2

′p

ΔD g + 21 ′ p

(7)

Where the first term in equation 7 represents the pressure-depth relationship for a standard ocean, and the second term in equation 7 is the contribution to depth due to localised density effects or the local deviation from the standard ocean, the so-called geopotential anomaly. The advantage of expressing the pressure depth relation in this way is that the first term only depends on pressure and latitude. The second term is considered of lower order and estimated to contribute from 0 m to 3 m at 5000 m depths, according to Saunders and Foffonov (1976). It should be noted that when this approximation was proposed, the pressure sensor accuracy (Baker et al., 1981) was lower than the contribution of this second corrective term, thus in many cases, it could be ignored (Saunders, 1981). Many subsequent works followed a similar approach. Saunders and Foffonov (1976) approximated the leading standard ocean term using the KnudsenEkman formulations for density and the international gravity formula (see Saunders and Foffonov 1976) – in essence a ‘first attempt’ at an approximation in the form of the UNESCO equation. Saunders (1981) improved on this work by incorporating the newly formulated density function termed ‘EOS-80’ (Equation of State 1980, see Foffonov and Millard, 1983). He then approximated the pressure-depth integral with a quadratic expression.

2.1 UNESCO pressure-to-depth approximation Foffonov and Millard (1983) derived a more accurate approximation to the first term of equation 7 by fitting a 4th order polynomial in pressure to a table of values computed from the exact expression based on the EOS-80 density formula. The table

02-SUT297-35(1)-ut66631.indd 5

3/27/18 1:50 PM

Puig. Errors in pressure to depth approximations and how to avoid them

included values in a range of 0–12 000 decibars and the accuracy of the fit was estimated to be equivalent to 0.0002 m or better (Foffonov and Millard, 1983). Please note this is a relative accuracy in relation to the test data and not an absolute measure of the approximation. Their polynomial fit is what is known as the UNESCO pressure-depth relationship: z( p, p )=

c1 p +c + c2 p 2 + c p 3 + c p 4 (g ( )+ ′ p )

(8)

where phi (ø) is the latitude in degrees. The accuracy of this approximation can be increased by resolving the second term in equation 7, the geopotential anomaly. For comparative purposes, this paper will refer to the two-term solution as the ‘UNESCO density’ solution (equation 7 and the approximation that ignores the geopotential anomaly term the ‘’UNESCO pressure’’ (equation 8). Leroy and Parthiot (1998) proposed a simple corrective term to be added to the UNESCO pressure approximation that enhances the accuracy by approximating the geopotential anomaly. They found that the corrective term approximation could account for most open-ocean situations within 0.8 m, what they termed the common ocean from 60°N to 40°S correction. Leroy (2007) later published an erratum, as typos in the corrective terms were discovered in the original publication. The Leroy and Parthiot corrective terms are not of wide use in operational hydrographic work and have been mentioned for completeness.

2.2. Average density approximation The simplest approximation to equation 5 is to assume that density does not depend on pressure, and replace it with the average density ( ρ∼) from surface to the measured pressure level: z≈

p 0 + 21 γ ′ p ) ρ(g

(9)

This is a linear relationship between pressure and depth that ignores sea water compressibility. Compressibility effects are described by the bulk modulus of sea water. It is estimated sea water compresses ~1 % in 1500 m of water (see Foffonov and Millard, 1983).

2.3. Numerical solution The following numerical integration method to equation 5 that is presented here eliminates the need for pressure-depth approximations as described earlier. This solution is then compared to the

different approximations to highlight their inaccuracies and limitations. Equation 5 can be solved using a numerical integration technique called ‘trapezoidal integration’. The steps to be followed are: 1. Conduct a CTP profile of the water column from surface to seabed. Ideally, a measurement should be taken every 1 decibars, approximately every 1 m for deep-water locations and perhaps at a smaller interval for shallow water sites. Process the CTP profile data as per WG51 (1991); 2. Compute profile density using the EOS-80 formula as per Foffonov and Millard (1983); 3. Compute local surface gravity using international gravity, according to latitude of the location (see Foffonov and Millard, 1983); 4. Use a suitable numerical integration scheme to compute the integral of density over pressure, like trapezoidal integration. Ideally, a table will be produced with a column for incremental water column pressure and a corresponding integrated density over pressure value for each pressure, for the given measured density profile‡. This method is suited to compute depths for a CTP profile. If depth at a certain pressure is required, the table produced in the 4 steps can be used to interpolate an integrated density value for the required pressure level.

3. Results 3.1. Veriﬁcation of the numerical algorithm Before any comparisons using real CTP profiles, the numerical solution described in section 2.3 was verified using literature data. Saunders (1981) is the only published reference for depth calculations using the hydrostatic equation based upon the standard ocean model. In this idealised ocean, the water column has a uniform profile temperature of 0°C, salinity of 35 on the Practical Salinity Scale 1978, at latitude of 30°, from sea level p = 0 to p = 1000 bar at maximum depth. Fig 1 reproduces the computations carried out by Saunders and those calculated in section 2. An additional numerical exercise was conducted to verify the effects caused by different sampling rates of the density profile. The numerical integration pressure increments were varied from 0.1, 0.5, 1 and 5 decibar increments. Saunders results were reproduced for all the profiles, with no significantly observed effect on the depth conversion accuracy between them. ‡ For calculations presented here see www.mathworks.com/help/matlab/ref/ trapz.html?requestedDomain=www.mathworks.com.

02-SUT297-35(1)-ut66631.indd 6

3/27/18 1:50 PM

Underwater Technology Vol. 35, No. 1, 2018

Depth-pressure relationship in a standard ocean, with T = 0, S = 35, latitude 30º, surface gravity 9.79324 m s−2. Depth (m) Pressure (db)

S and F*

500 1000 1500 2000 2500 3000 4000 5000 7500 1000

495.99 990.77 1484.40 1976.86 2468.18 2958.38 3935.49 4908.31 7322.36 9712.02

This paper

Pressure (decibar)

dp = 1 decibar Z(m)

Quadratic error

496.00 990.81 1484.44 1976.96 2468.27 2958.50 3935.65 4908.51 7322.65 9712.45

500

496.00

1000

990.81

1500

1484.45

2000

1976.96

2500

2468.27

3000

2958.50

4000

3935.65

5000

4908.51

7500

7322.65

10000

9712.45

0.17 0.25 0.27 0.23 0.17 0.09 –0.06 –0.08 1.20 5.77

*Table 1 of Saunders and Fofonoff.

Fig 1: Depth-pressure relationship for a standard ocean, taken from Saunders (1981), and comparison to numerical solution of hydrostatic equation with density sampled at 1 decibar interval

Table 1: Summary of CTD proﬁles Location

Obs date

Lat (Deg)

Lon (Deg)

Max depth (m)

Average S (psu)

Average T (°C)

Average density (Kg/m3)

East Med SW Atlantic Bay of Bengal Norwegian Shelf Mid-East Atlantic SE Atlantic South China Sea Timor Sea NW Australian Shelf North Sea

07/2004 09/2010 03/2008 04/2004 07/2004 07/2006 08/2012 12/2011 03/2007 07/2008

32N 24S 16N 65N 43N 07S 06N 12S 19S 56N

030E 044W 082E 006E 010W 011E 114E 124E 116E 003E

935 1270 867 1024 3824 1286 1237 77 132 123

39 34.5 37.9 35.1 36.1 34.8 34.5 34.2 34.9 35.1

14.7 8.4 11.1 –1.6 4.4 6.7 7.7 29.0 24.8 7.8

1031.07 1029.61 1030.68 1030.41 1037.12 1030.06 1029.39 1021.59 1023.58 1027.56

3.2. Complete proﬁle depth comparisons Ten CTP profiles were selected from data gathered by the author over a period of eight years of work in offshore projects around the world. The data in Table 1 summarise their characteristics. All profile data were processed the same way, the CTD was tared§ at surface, thus negating atmospheric pressure. No filters or smoothing were applied; the data were only cleaned for blunders and repetitive data. The profiles were arranged in monotonic pressure increments from surface to maximum pressure reading. The profiles considered consist of conductivity, temperature and pressure or salinity. The focus is not the processing of the CTP data or its accuracy, rather the differences in depth computed from pressure from the different depth approximations. It is worth mentioning that the focus of the paper is pressure-to-depth computation in a hydrographic § This term refers to the procedure of removing the atmospheric pressure at sea level.

operational environment hence not all oceanic applications are represented in this study of limited scope, with specific emphasis on full oceanic depth applications. Depth was computed for every profile using the numerical integration scheme and approximation formulas described, except that of Leroy and Parthiot (1998) who are not widely used in operational hydrography. Computed density for each profile is presented in Fig 1. The first schematic shows the complete profiles down to depth. The CTD profile taken off the Spanish coast extends down to 3800 m, which is considerably deeper than the rest, thus a close-up at 1250 m is presented within with Fig 2. Depth was computed for each discrete pressure increment, for all profiles, using each of the depth formulae described in section 2. A residual was computed for each approximation relative to the numerical solution presented in section 2. The root-mean-square (rms) of the residuals

02-SUT297-35(1)-ut66631.indd 7

3/27/18 1:50 PM

Puig. Errors in pressure to depth approximations and how to avoid them

1020 0

Density in kg/m3 1030 1035 1040

1025

1045

1050 East Mediterranean SW Atlantic Bay of Bengal North Atlantic Mid E Atlantic SE Atlantic South China Sea Timor Sea Australian NW Shelf North Sea, Norwegian Sector

500

1000

2000

1032

1034

1030

1028

1026

1024

1020

Density in kg/m3 1022

Depth in metres from sea level

1500

2500

3500

Depth in metres from sea level

3000

200

400

600

800

1000

1200

4000

Fig 2: Density proﬁles for all CTD casts used in comparisons. The inset is a close-up of the ﬁrst 1500 m of water depth

was then calculated for each approximation in all ten CTP profiles. The results are summarised in Table 2. From Table 2 it can be observed that the simple pressure-based approximations experience the largest residual. The depth as computed from the average density formula is very close to the numerical solution; the difference increases with depth, and in general it is comparable to the UNESCO density approximation. The biggest

difference is for the deepest profile, where maximum depth reaches 3800 m and a maximum difference of ~60 cm is apparent when compared to the numerical solution.

3.3. Seabed proﬁle depth comparisons The majority of subsea work is carried out on the seabed and in a layer 10 m above it, therefore it was worthwhile to investigate the accuracy of the approximations in this layer alone. To do this, the

02-SUT297-35(1)-ut66631.indd 8

3/27/18 1:50 PM

Underwater Technology Vol. 35, No. 1, 2018

Table 2: Computed root-mean-square (rms) for differences between approximate depth computations and the numerical solution for the complete profiles from surface to seabed/ maximum depth. Average over the profiles is also calculated Profile

East Med SW Atlantic Bay of Bengal Norwegian Shelf Mid-East Atlantic SE Atlantic South China Sea Timor Sea NW Australian Shelf North Sea Computed average rms (m)

UNESCO pressure (m)

UNESCO density (m)

Average density solution (m)

0.35 1.31 0.26 0.25 0.82 0.94 1.52 0.31 0.39

0.01 0.03 0.05 0.02 0.02 0.01 0.05 0.04 0.01

0.04 0.06 0.04 0.01 0.67 0.07 0.05 0.00 0.00

0.09 0.62

0.00 0.02

0.01 0.09

Table 3: Computed root-mean-square (rms) for differences between approximate depth computations and the numerical solution for the last 10 m of the profiles from seabed/ maximum depth upwards. Average over the profiles is also calculated Profile

UNESCO pressure (m)

UNESCO density (m)

Average density solution (m)

East Med SW Atlantic Bay of Bengal Norwegian Shelf Mid-East Atlantic SE Atlantic South China Sea Timor Sea NW Australian Shelf North Sea Computed average rms (m)

0.67 –1.91 0.32 –0.27 –0.03 –1.40 –2.06 –0.48 –0.60 –0.11 1.05

–0.01 –0.03 –0.05 –0.01 0.00 –0.01 –0.05 –0.04 –0.01 0.00 0.03

–0.10 –0.14 –0.02 –0.01 1.51 0.16 0.11 0.00 0.00 –0.01 0.48

Table 4: Computed relative depth differences simulating metrology hub-to-hub height difference of 5 m for 5 of the 10 CTD proﬁles. UP is ‘UNESCO pressure’, ‘UD is ‘UNESCO density’, and AD is ‘average density’ computed depths Eastern Med (max depth of proﬁle 935 m) Pressure (decibar)

Numerical (m)

UP (m)

UD (m)

AD (m)

942.375 937.375

932.995 928.057

933.671 928.729

932.989 928.050

933.103 928.160

4.939

4.943 –0.004

4.939 0.000

4.943 –0.005

UP (m) 1266.694 1261.944

UD (m) 1268.577 1263.824

AD (m) 1268.564 1263.811

4.750 0.003

4.753 0.000

Difference (m) Difference from numerical (m)

SW Atlantic (max depth of proﬁle 1270 m) Pressure (decibar) 1278.748 1273.938 Difference (m) Difference from numerical (m)

Numerical (m) 1268.604 1263.850 4.753

Bay of Bengal (max depth of proﬁle 867 m) Pressure (decibar)

Numerical (m)

UP (m)

UD (m)

AD (m)

872.060 866.842

864.758 859.599

865.080 859.915

864.706 859.547

864.738 859.578

5.159

5.165 –0.006

5.159 0.000

5.160 –0.001

Difference (m) Difference from numerical (m)

Norwegian Shelf (max depth of proﬁle 1024 m) Pressure (decibar)

Numerical (m)

UP (m)

UD (m)

AD (m)

1035.337 1030.995

1022.913 1018.635

1022.641 1018.362

1022.899 1018.621

1022.905 1018.624

4.278

4.278 0.000

4.281 –0.003

Difference (m) Difference from numerical (m)

Mid-Eastern Atlantic (max depth of proﬁle 3824 m) Pressure (decibar)

Numerical (m)

UP (m)

UD (m)

AD (m)

3887.609 3882.782

3821.823 3817.124

3821.800 3817.097

3821.823 3817.124

3823.333 3818.631

4.699

4.703 –0.004

4.699 0.000

4.702 –0.004

Difference (m) Difference from numerical (m)

02-SUT297-35(1)-ut66631.indd 9

3/27/18 1:50 PM

Puig. Errors in pressure to depth approximations and how to avoid them

–1.20 0

–1.00

Computed depth residual (m) –0.80 –0.60 –0.40 –0.20

0.00

0.20

UNESCO pressure UNESCO density Average density 500

1000

Depth (m)

1500

2000

2500

3000

3500

4000

Fig 3: Mid-East Atlantic CTD proﬁle computed depth residual for the different approximations compared to the numerical solution of the hydrostatic pressure equation. The vertical axis is proﬁle depth in metres

average of the residuals for each approximation was computed for a water column layer 10 m thick from maximum depth upwards of the assumed seabed. Table 3 summarises these results. The computed residuals for the last 10 m of the water column are larger when compared to those computed for the complete profiles. This is no surprise given this is maximum pressure for the profiles and, in some cases, compressibility of seawater and complex density structures have a greater impact. From Table 3 it can be observed in general that the UNESCO pressure approximation underestimates the depth for most ocean basins, except where the profiles have high salinity averages.

These residuals are in the order of 1 m, with a maximum of 2 m, in accordance with the literature and stated accuracy of the approximation (Foffonov and Millard, 1983). This behaviour can also be seen in Fig 3 for the deepest CTP profile in the test cases. Adding the geopotential anomaly correction or the UNESCO density equates the estimate of depth to that of the numerical solution, with centimetre differences observed. The average density formula as observed earlier is found to be in good agreement with the numerical solution; however, it seems to increase with depth to a maximum of 1.5 m at 3800 m depth.

02-SUT297-35(1)-ut66631.indd 10

3/27/18 1:50 PM

Underwater Technology Vol. 35, No. 1, 2018

Table 5: Computed relative depth differences simulating metrology hub-to-hub height difference of 5 m for the remaining CTD proﬁles. UP is ‘UNESCO pressure’, UD is ‘UNESCO density’ and AD is ‘average density’ computed depths. South East Atlantic (max depth 1286 m) Pressure (decibar)

Numerical (m)

UP (m)

UD (m)

AD (m)

1294.156 1289.086

1284.337 1279.323

1282.935 1277.924

1284.329 1279.316

1284.493 1279.478

5.014

5.011 0.003

5.014 0.000

5.015 –0.002

Difference (m) Difference from numerical (m)

South China Sea (max depth 1237 m) Pressure (decibar)

Numerical (m)

UP (m)

UD (m)

AD (m)

1243.991 1238.978

1235.459 1230.500

1233.401 1228.445

1235.412 1230.453

1235.560 1230.619

4.959

4.956 0.003

4.959 0.000

4.941 0.018

Difference (m) Difference from numerical (m) Timor Sea (max depth 77 m) Pressure (decibar)

Numerical (m)

UP (m)

UD (m)

AD (m)

75.012 70.024

75.062 70.076

74.567 69.610

75.026 70.040

75.059 70.074

4.986

4.958 0.028

4.986 0.000

4.985 0.000

Difference (m) Difference from numerical (m)

North West Australian Shelf (max depth 132 m) Pressure (decibar)

Numerical (m)

UP (m)

UD (m)

AD (m)

128.920 123.940

128.706 123.742

128.094 123.148

128.694 123.730

128.704 123.740

4.964

4.946 0.018

4.964 0.000

Difference (m) Difference from numerical (m) North Sea (max depth 123 m) Pressure (decibar)

Numerical (m)

UP (m)

UD (m)

AD (m)

121.900 115.900

120.866 114.922

120.752 114.810

120.862 114.918

120.854 114.910

5.944

5.942 0.002

5.944 0.000

5.943 0.001

Difference (m) Difference from numerical (m)

3.4. Relative differences Up to now, this paper has investigated only absolute accuracy of the approximations, but as mentioned previously relative depth accuracy is also important. To investigate what the implication of using approximations for these types of measurements is, two pressure levels are selected from each of the ten profiles: one ~2 m up from deepest reading, assumed to be seabed; and then another ~7 m up in order to make the depth difference ~5 m, a typical metrology hub-to-hub difference. The difference in relative depth computed from the numerical solution is then compared to each relative depth computed using each of the approximations for pressure depth conversion. The summary of these relative depth computations is shown in Tables 4 and 5. These tables show that, even in the worst cases (Timor Sea for the UNESCO pressure approximation), using any approximation

results in a difference to the numerical solution that is lower than the typical accuracy for a pressure sensor unit and under typical metrology accuracy requirements.

4. Discussion A summary of the most common approximations for the pressure-to-depth approximation have been presented, together with a numerical integration method of the hydrostatic equation. Depth computation comparisons have been made between the numerical integration formula and commonly used approximations using CTP cast data from different ocean regions. For absolute depth measurements, using the UNESCO pressure formula introduces potential errors that surpass the apparent accuracy offered by modern pressure sensors, except in those locations

02-SUT297-35(1)-ut66631.indd 11

3/27/18 1:50 PM

Puig. Errors in pressure to depth approximations and how to avoid them

whose environmental characteristics resemble those of a standard ocean. The error is generally observed to increase with depth and in the tropical latitudes. The addition of the geopotential anomaly corrective term (UNESCO density) to the UNESCO pressure depth computation reduces the observed errors with the pressure-based approximation. The errors observed relative to the numerical solution for the UNESCO density approximation are within the centimetre range and probably due to numerical computation effects of both schemes. The average density formula works well given its simple nature, and this is due to the quasi-linear nature of the pressure-density relationship in shallow water. However, its accuracy degrades as the relationship between pressure and density becomes nonlinear with depth.

5. Conclusion Errors in pressure-based depth approximation solutions are now typically larger than the errors of pressure sensors. The wealth of available density data on modern offshore survey packages should be used to compute depth that uses the full potential of quartz-based pressure sensor accuracies. The numerical integration of the hydrostatic pressure equation yields the same depth computations as the UNESCO density formula and can be considered a more elegant solution. However, it is a good operational practice to check all density derived depth computations against a UNESCO pressure computation, as it proves to be a great gross error check for density measurement problems. The average density formula should be used with caution and within the limits imposed by its approximation assumptions. For relative depth computation as discussed here, the errors are typically smaller than pressure sensor accuracy. This is related to the small range of pressure change and the high resolution exhibited by quartz pressure sensors. Interesting follow-up research would be to study the implications of using more complex gravity models and measured data in the hydrostatic depth solution. In addition, follow-up research could study the implications of using the new TEOS-10 thermodynamic equations to estimate

density from CTD profiles and the implementation of real-time dynamic pressure accountability using global circulation models and near-real time observed data.

References Baker DJ. (1981). Ocean Instruments and Experiment Design. In: Warren BA and Wunsch C (eds). Evolution of Physical Oceanography. Cambridge: MIT Press, 396–433 pp. Barret S, Canning S, Prytz F, Puig J and Vickery K. (2017). Guidance on Subsea Metrology. IMCA Survey Division, September 2017. IMCAS019. Doornhof D, Kristiansen TG, Nagel NB, Patillo PD and Sayers C. (2006). Compaction and subsidence. Oilfield Review, Autum 2006: 50–68. Foffonov JP and Millard RC. (1983). Algorithms for Computation of fundamental properties of sea water. UNESCO Technical Pappers in Marine Science. Number 44. Intergovernmental Oceanographic Commission (IOC). (2010). The international thermodynamic equation of seawater – 2010: Calculation and use of thermodynamic properties. Manuals and Guides Number 56. UNESCO. 196 pp. LaCasce JH. (2013). Atmosphere-Ocean Dynamics. Norway: University of Oslo. Available at: http://www.uio.no/studier/ emner/matnat/geofag/GEF4500/h13/gef4500jhl6.pdf <last accessed on 22/02/18>. Leroy CC. (2007). Erratum: Depth-Pressure relationship in the oceans and sea. Journal of the Acoustic Society of America 121: 2447. Leroy CC and Parthiot F. (1998). Depth-Pressure relationship in the oceans and sea. Journal of the Acoustic Society of America 103: 1346–1352. Paroscientific. (n.d.). The Advantages of Quartz Technology. Version 1.0.0.0. Available at: http://paroscientific. com/pdf/400%20Resonant%20Quartz%20Crystal%20 Technology.pdf <last accessed 02/02/18>. Paros JM and Kobayashi T. (n.d.). Calibration methods to eliminate Quatyz Sensor Drift. Technical Note G8097 Rev F. Redmond WA, USA: Paroscientific Inc. Pawlowicz R. (2013). What every oceanographer should know about TEOS-10. The TEOS-10 Primer v8. Available at: www.teos10.org/pubs/TEOS-10_Primer.pdf <last accessed 2/2/18>. Pickard GL and Emery W. (1990). Descriptive Physical Oceanography, An Introduction. Oxford: Pergamon. 320 pp. Resnick R, Halliday D and Krane K. (2001). Physics, Vol 1 5th edition. Hoboken: John Wiley and Sons, 624 pp. Saunders PM. Practial Conversion of Pressure to Depth. (1981). Journal of Physical Oceanography 11: 573–574. Saunders PM and Foffonov JP. (1976). Conversion of pressure to depth in the Ocean. Deep Sea Research 23: 109–111. WG51, SCOR. The Acquisition, calibration and analysis of CTD data. UNESCO Technical Pappers in Marine Science, Volume 54, 1991. [Online January 2018] http://earth-info. nga.mil/GandG/publications/tr8350.2/wgs84fin.pdf.

02-SUT297-35(1)-ut66631.indd 12

3/27/18 1:50 PM

doi:10.3723/ut.35.013 Underwater Technology, Vol. 35, No. 1, pp. 13–22, 2018

Technical Paper

www.sut.org

Design and implementation of a remotely operated vehicle testbed Zhigang Deng*, Fang Yuan, Daqi Zhu and Feng Xue Laboratory of Underwater Vehicles and Intelligent Systems, Shanghai Maritime University, Shanghai, 201306, China Received 25 October 2016; Accepted 29 November 2017

Abstract A SMU-II ROV has been designed to highlight some of the characteristics of engineering applications for underwater vehicles, such as continuous working hours, simple crawling function, and perception of the environment. Its development is based on the SMU-I autonomous and remotely operated vehicle (ARV), with improvements mainly in the following aspects: a cable power supply instead of a battery-powered supply to extend the underwater operating time; a simple manipulator to grab underwater objects; ranging sonar for acoustic location and measurement of the echo characteristics of ‘targets’ in the water; and a dual camera to replace single camera that can help to carry out follow-up binocular vision research. The main improvement is that the manipulator uses a rotator unit with the special design of a hand-like claw, allowing it to easily grab irregular objects. The fuzzy proportional-integral-derivative (PID) control method is adopted, and the experimental results show that the comprehensive performance can meet the underwater motion control and operation requirements. Keywords: programmable logic controller (PLC), remotelyoperated vehicle (ROV), test-bed, user data protocol (UDP), VxWorks

1. Introduction Today, there are mainly three kinds of unmanned underwater vehicles (UUV): remotely operated vehicles (ROVs – Behar et al., 2015; Anwar et al., 2016), autonomous underwater vehicles (AUV – Fischer et al., 2014; Zain et al., 2015) and underwater gliders (Page et al., 2017; Singh et al., 2017). A great array of ROVs and AUVs has been produced – such as biorobotic AUV (Bandyopadhyay, 2005), deep-water ROV (Salgado-Jimenez et al., 2010; Ramadass et al., 2015) and long-range, high-speed

* Corresponding author. Email address: dzg1026@126.com

AUV (Shen et al., 2016) – thus enhancing the range and performance of UUVs. On the basis of SMU-I autonomous and remotely operated vehicle (ARV – Deng et al., 2014), SMU-II ROV is designed to act as a test-bed platform for experimental validation of control algorithms and binocular vision systems (Chen et al., 2015; NadalSerrano and Lopez-Vallejo, 2015; Javdani et al., 2016). The design criteria of the vehicle are defined as follows: 1. The use of flat semi-streamlined structure to get drag reduction; 2. Power supply from water surface to enhance endurance capability; 3. Single electronic cabin instead of double electronic cabin for easy manufacture and maintenance. The development of SMU-II is described in this paper. Compared with SMU-I, SMU-II makes improvements mainly in the following aspects: a cable power supply is used instead of a lithium battery-powered supply to extend the underwater operating time; a simple manipulator is designed to grab irregular underwater objects; ranging sonar is used as a means of acoustic location and measurement of the echo characteristics of ‘targets’ in the water; a dual camera replaces a single camera to carry out follow-up binocular vision research. This ROV consists of a mechanical and an electrical system, along with the integration of subsystems including various sensors (such as sonar, rate gyroscope, vision, compass, depth, and current and voltage sensors). The ROV is designed as a test-bed platform for a variety of research in underwater technologies especially involving high-performance and multipurpose ROV. In this paper, the specifications of the SMU-II’s hardware and software design are discussed, and

03-SUT277-35(1)-ut66627.indd 13

3/28/18 4:25 PM

Deng et al. Design and implementation of a remotely operated vehicle testbed

experiment results are shown. Section 2 introduces the design and development of its mechanical and electrical system. The integration of subsystems is covered as well. Section 3 describes the communication protocol and control design, while the tank test is presented in section 4. Section 5 gives some concluding remarks and future work prospects on the vehicle.

2. Vehicle description 2.1. Overview SMU-II is an open-framed ROV that is 1.4 m × 0.63 m × 0.46 m (length, width and height), weighs 60 kg, and has a maximum underwater depth rating of 150 m (see Fig 1). The specifications of the vehicle and its main electrical components are illustrated in Tables 1 and 2, respectively. Table 3 presents the properties of the sensor signals of the control system. The ROV possesses good manoeuvrability owing to 4 degrees of freedom (DOFs), including surge, heave, roll, and yaw. The vehicle is designed to be passively stable in roll and pitch (Jiang et al., 2000). The hardware architecture design of the vehicle, shown in Fig 2, consists of two parts. The surface control unit (SCU – see Fig 2(a)) is mainly used to transmit operating instructions and to receive underwater system information. It comprises an industrial computer (GEN-E9455), a programmable logic controller (WAGO 750-881), and a power unit (SP480-48/PBA100F-24). The main components include, among others, joysticks, a monitor and an video capture card. The underwater system (see Fig 2(b)), mainly used to receive operating instructions and transmit underwater system information, comprises a Intel Pentium M processor board (PMI-6D), a multiserial-port board (MSP-12), a data acquisition board (DMM-32X) and a power board HE104. The main components include sonar, pressure sensor, camera, angular rate sensor and thrusters. 2.2. Mechanical design The schematic design of SMU-II ROV is shown in Fig 3. The vehicle is 1.4 m long and 0.63 m wide, and it has been designed to include a frame structure for modification of mounted devices and thruster allocation. The ROV has one hull, which is 0.57 m in length, 0.18 m in diameter and 0.006 m thick. The pressure hull provides the essential buoyancy and the watertight compartment for all on-board electronics and sensors. Two caps are especially designed to complete the hull and are attached to each end so that they reliably seal the hull. Sealing

1: Thrusters; 2: Aluminium frames; 3: Thruster; 4: Camera; 5: LED lamp; 6: Manipulator; 7: Sonar; 8: Buoyancy materials; 9: Neutrally buoyant tether

Fig 1: Physical layout of the vehicle

Table 1: Speciﬁcations of the ROV Dimensions Depth rating Weight Maximal speed Buoyancy materials

(1.4 × 0.63 × 0.46) m 150 m 60 kg 4.0 knot Epoxy resin and hollow glass microspheres 2 horizontal and 2 vertical Digital compass Pressure sensor (depth sensor) Ranging sonar Angular rate sensor 2 1/3″ DSP CCD LED lights

Thrusters Sensors

Video cameras Lighting

Table 2: Speciﬁcations of main electrical components Part name

Type

Manufacturer

Thrusters Ranging sonar

Model:260 852

Outland Technology Imagenex Technology Corp.

LED Video cameras Intel Pentium M processor board Industrial computer Programmable logic controller

LED-G UWV325 PMI-6D

Teledyne Bowtech Outland Technology SBS Science & Technology Co., Ltd. AAEON WAGO Kontakttechnik GmbH & Co. KG

GEN-E9455 750-881

is achieved with commercially available O-rings. The front cap has three watertight connectors, while the rear cap has five watertight connectors for electric connection. The caps also make the interior accessible for easy repair and maintenance in the field. The yellow cap on top of the vehicle is a buoyant material with density of 0.28 × 103 kg/m3, and the buoyancy of the vehicle in water and gravity is almost equal, providing an overall slightly floating state.

03-SUT277-35(1)-ut66627.indd 14

3/28/18 4:25 PM

Underwater Technology Vol. 35, No. 1, 2018

Table 3: Sensor properties Variable

Sensor (manufacturer)

Precision

Update rate

Output

Heading Roll and pitch Depth Angular rate

Compass, HMR3000 (Honeywell) Compass, HMR3000 (Honeywell) Pressure sensor, YB-KO-LVG (Honeywell) Gyro, Crs03-02 (Silicon)

<0.5° ±0.4° 5.0 cm Bias<±1°/s

10 Hz 10 Hz 20 Hz 10 Hz

Digital Digital Analogue Analogue

(a)

GEN-E9455

Joysticks

WAGO750-841 Monitor

PBA100F-24 Video capture card

SP-480-48 Master controller

(b) Camera Sonar Slave controller PMI-6D-N Angular rate sensor

Compass

MSP-12-N DMM-32X-AT

Thrusters Manipulator

HE104 Pressure sensor

Fig 2: Hardware architecture design of the SMU-II ROV

(a) Top

(b) Right Fig 3: Schematic design of the vehicle (unit: mm)

03-SUT277-35(1)-ut66627.indd 15

3/28/18 4:25 PM

Deng et al. Design and implementation of a remotely operated vehicle testbed

The propulsion system for this project consists of four ducted propellers. The propellers are driven by 150 voltage of direct current (VDC) motors. To provide both forward and backward movement, two horizontal thrusters are mounted aft of the vehicle. Yaw is provided by operating the thrusters in opposite directions. Two vertical thrusters are mounted symmetrically on the right and left sides of the vehicle, which makes the vehicleâ&#x20AC;&#x2122;s motion in heave and roll possible.

2.3. Control system architecture As the core part of the ROV, the control system is directly related to its system performance, work capacity and running environment. The entire control system has several subsystems for surface display and control, motion control implementation, sensors, power and cable transmission. The overall architecture of the ROV control system is illustrated in Fig 4. Power supply for the underwater system is 150 VDC (4 SP-480-48 modules), which makes it easy to power up the vehicle. Through a thin neutrally buoyant tether, video signal is transmitted directly to the surface monitor. In the underwater system, to deal with data acquisition, information

processing and output calculating, an extremely low-power Intel Pentium M processor board is used, as well as VxWorks real-time operating system (Zhang and Yu, 2011). The SCU, whose power supply is 220 voltage of alternating current (ACV), uses a high-performance, low-power Intel AtomTM N270 processor to process the information on interactions between the underwater system and the SCU. The operating system, Windows XP, has good application compatibility. In the SMU-II ROV, the underwater system is the most important part of the whole control system and consists of several main boards. The Pentium M PC/104 module (PMI-6D) is a highly integrated central processing unit (CPU) board solution for embedded system designers, featuring the compact form factor, a wide operating temperature range, and low consumption. It can be stacked with other PC/104 expansion modules to form a highly integrated control system. PMI-6D is a high-density performance board with: a keyboard and mouse interface; two serial ports and one parallel port; 10/100 Base-T Ethernet; a compact flash (CF) card socket on-board; an enhanced integrated drive electronics (EIDE) interface; four USB 2.0 ports; a USB boot capability/legacy; and an Intel 220V AC

MCU board GEN E9455

Potentiometers joystick

AC/DC 220~5

Push buttons

AC/DC 220~24

LCD monitor

AC/DC 220~12

WAGO 750-881

AC/DC 220~150 Surface control unit

150V DC Camera Sonar

DC/DC 150-24

Manipulator PMI-6D-N MSP-12-N DMM-32X-AT

Water alarm Pressure sensor

DC/DC 150-5

Digital compass Angular rate sensor DC motor

DC/DC 150-12

Underwater system Control signal

Sensor data

Power supply

Sonar/video

Fig 4: Overall architecture of the ROV control system

03-SUT277-35(1)-ut66627.indd 16

3/28/18 4:25 PM

Underwater Technology Vol. 35, No. 1, 2018

integrated graphics controller with dual independent display support and dual-channel low-voltage differential signalling (LVDS) interface. The MSP-12 is a multi-serial port board with 12 peripheral component interconnect (PCI) serial ports with 128 byte first input, first output (FIFO); the RS232/RS485/RS422 mode is optional. DMM32X-AT is a PC/104-format data acquisition board with a full set of analogue and digital I/O features. It offers 32 analogue inputs with 16-bit resolution and a programmable input range; 250 000 samples per second maximum sampling rate with FIFO operation; 4 analogue outputs with 12-bit resolution; user-adjustable analogue output ranges; 31 lines of digital I/O; one 32-bit counter/timer for the analogueto-digital converter (ADC) and interrupt timing; and one 16-bit counter/timer for general purpose use. All signals are connected to the serial board and data acquisition board.

1: DC-DC; 2: Power distribution board; 3: Water-leakage alarm; 4: Angular rate sensor; 5: MCU board; 6: Voltage sensors; 7: Digital compass; 8: Watertight connector

Fig 5: Underwater system

2.3.1. Underwater system The structure and arrangement of mounted devices are shown in Fig 5. They are designed to be installed in the pressure hull. The state sensors include voltage and current sensors and water-leakage alarm. The motion sensors include a forward-looking ranging sonar, a pressure sensor, angular rate sensor and a digital compass. The scanning sonar, 852 Ultra-Miniature, can scan a maximum of 360° with a maximum range of 50 m. The angular rate sensor, Crs03-02, gives advanced and stable performance over time and in a range of temperatures, overcoming the mount sensitivity problems experienced with simple-beam or turning fork-based sensors. The digital compass, HMR3000, which can provide accurate heading (0.1°) measurement over a 0°–360° tilt range, is mainly used for measuring heading. Its working voltage is 5 VDC with sampling frequency 20 Hz, and a RS232 serial data port is used for connection to the microcontroller unit (MCU) board. It also provides an accurate angular (0.1°) measurement over a ±40° tilt range of roll and pitch. The pressure sensor can output 0 V–5 V voltage signal. All the sensor data above can be acquired by the underwater MCU. Two charge coupled device (CCD) cameras, UWV325, provide a good view of the underwater environment to the operator. As shown in Fig 1, LED lamps and thrusters are attached to the vehicle frame. They are all driven by voltage signal. The maximum output of a thruster is 300 W with a supply voltage of 150 V. According to the surface commands, these actuators update their status and action. A flowchart of the underwater system is shown in Fig 6.

Voltage A/D

Pressure Angular rate

Surface control unit

D/A

Thrusters

Power switches

Water alarm

RS232

Compass

RS485

Manipulator

RS485

Sonar

RS485 PMI-8N

Fig 6: Flowchart of underwater system

03-SUT277-35(1)-ut66627.indd 17

3/28/18 4:25 PM

Deng et al. Design and implementation of a remotely operated vehicle testbed

2.3.2. Surface control unit The whole ROV system can be operated by the SCU, as shown in Fig 7. A thin umbilical cable connects SCU to the vehicle for providing data and a video link between them. The PLC module receives the control commands, which is sent out by joysticks and push buttons, and sends to the industrial computer (GEN-E9455) via user data protocol (UDP). The sensor data and underwater images are displayed on the LCD monitor in real time. A data flow diagram of surface control unit is

1: LCD monitor; 2: Mini-keyboard; 3: Potentiometer (vertical); 4: Neutral tether; 5: Joystick (horizontal); 6: Power switch; 7: Control switch; 8:Power cable

Fig 7: Surface control unit

To the underwater system RS485

750-881 PLC

UDP

Ethernet switches UDP

Video capture card

GEN-E9455

Sonar

Fig 8: Data ﬂow diagram of the surface control unit

shown in Fig 8. ‘AI’ and ‘DI’ in Fig 8 indicate the analogue and digital inputs, respectively. The PLC 750-881 module is used to process the data from one analogue input module (750-457), two digital input modules (750-430) and an interrupt management module (750-600) during manual operation. The UDP is selected as a transport protocol. Only synchronous calls of the dynamic link library (DLL) are supported for calls from Visual Basic. The DLL that supports synchronous and asynchronous reading and writing of values, and specifications of Modbus TCP’s MBT read registers are shown in Table 4.

2.4. Software The SMU-II’s software architecture (shown in Fig 9) is designed to be a scalable, flexible, modular, reliable and compatible structure so that it can handle any modification or additions. The real-time operating system VxWorks is embedded in PMI MCU. VxWorks is known for its high reliability and realtime performance, and has a wide range of applications in industries such as aviation, military and medicine. The heart of the VxWorks run-time system is the highly efficient wind microkernel. The target development has features such as dynamic linking loader and flexible booting from read-only memory (ROM), local disk or over the network. Both the SCU and the embedded software comprise four universal modules. The Main Entry module is the entrance of the whole program. The Target Init module initialises on-board resources such as digital input output (DIO), ADC and digital-to-analogue converter (DAC). The ParaInit module initialises global variables and initial state for mounted devices. The User Task Init module creates user tasks. As shown in Fig 9, five tasks are designed for underwater software and three for surface software. The ‘communication’ task implements communication protocol and exchange data between surface and underwater. The ‘vehicle operation’ task is in charge of updating user operation and generating control commands. The ‘sensor sampling’ task carries out sensor data acquisition. Other devices such as lights and cameras are operated in

Table 4: Speciﬁcations of MBT read registers Word address (hex)

Read/write

Functions

Word values

00 01 02 03 04

R R R R R

Joystick-X axial Joystick-Y axial Thruster-vertical Pitch-tune Switch state

0×000=0V;0×1fff=neutral;0×3fff=5V 0×000=0V;0×1fff=neutral;0×3fff=5V 0×000=0V;0×1fff=neutral;0×3fff=5V 0×000=0V;0×1fff=neutral;0×3fff=5V Bit(0):lamp; Bit(1):grab; Bit(2):release; Bit(3):camera; Bit(4):auto-heading; Bit(5):sonar; Bit(6):auto-depth; Bit(7):power; Bit(8):auto-height; Bit(9): auto-speed

03-SUT277-35(1)-ut66627.indd 18

3/28/18 4:25 PM

Underwater Technology Vol. 35, No. 1, 2018

Windows XP

Target init

Vehicle operation task

Para init

Joystick press buttons

User task init

NULL Reserved task

Communication task

User tasks

Main entry

Moudle

Surface software

RS485 Lamp

NULL

Communication task

Sonar, heading, depth Sensor data task

Auto (heading, depth) Motion control task

Sonar, heading, depth

Main entry

Target init

User tasks

Reserved task

camera, motors

Para init

User task init

Moudle

Peripheral control task

Vxworks

Underwater software

Fig 9: Software architecture

Fuzzy logic inference ΔkP r

ΔkI

ΔkD

× –

PID controller

de dt

u Thrusters model

TCM

Vehicle dynamic

ec Sensor

Fig 10: Fuzzy-PID controller for depth control and heading control

the ‘peripheral control’ task. Control strategy, such as auto-heading and obstacle-avoidance, is done via the ‘motion control’ task. This architecture is effective in its performance of controllability, flexibility and stability.

3. Communication protocol and control design 3.1. Communication protocol The communication of the surface and underwater systems of the SMU-II is realised with the help of a thin umbilical cable through RS485, whose

band rate is 115 200 bps (bit per second). The underwater system has features with dynamic linking loader and flexible booting over the network, so the underwater system program can be updated through the surface system. The communication data format from underwater system to surface is shown is Table 5. The communication data format from surface system to underwater system is shown in Table 6.

3.2 Control design The SMU-II is designed to have 4 DOF including surge, yaw, roll and heave with the specific thruster 19

03-SUT277-35(1)-ut66627.indd 19

3/28/18 4:25 PM

Deng et al. Design and implementation of a remotely operated vehicle testbed

configuration. The kinematic and dynamic model can be expressed as equation (1): ⎪⎧⎪M (v ) + C (v )v + D (v )v + g ( ) = τ ⎨ ⎪⎪η = J ( )v ⎩

(1)

T is used to describe the forces and moments acting on the vehicle in the body-fixed frame. A simplified relationship between τ (the vector control inputs) and T can be expressed through the linear mapping: τ = LT∙ L is a matrix known as the thruster control matrix (TCM). The SMU-II has the following TCM:

where: M v C(v) D(v) g(η)

τ η J(η)

– the inertia matrix, including added mass; – the linear and angular velocity vector with coordinates in the body-fixed frame; – the matrix of Coriolis and centrifugal terms, including added mass; – the hydrodynamic damping matrix; – the vector of gravitational and buoyant forces and moments; – the vector of control inputs; – the position and orientation vector with coordinates in the earth-fixed frame, – a transformation matrix and the kinematics transformation relating the body-fixed velocities to the time derivative of the positions in the local geographical frame.

Table 5: Data format from underwater to surface Byte

Deﬁnition

Description

0–2 3–4 5–6 7–8 9–10 11–12 13–14 15–16 17–18 19–20 21–22 23–24 25–26 27–28 29 30

0×52, 0×4F, 0×56 Surge displacement Sway displacement Depth Heading angular Pitch angular Roll angular Surge speed Angular velocity z-axis Heave speed Power voltage(150V) NULL NULL NULL Water alarm state Check sum

ROV Low and high bytes Low and high bytes Low and high bytes Low and high bytes Low and high bytes Low and high bytes Low and high bytes Low and high bytes Low and high bytes Low and high bytes Spare Spare Spare Bit0 From byte 3 to byte 29

Table 6: Data format from surface to underwater Byte

Deﬁnition

Description

0–2 3

0×52, 0×4F, 0×56 Power switch

4 5–6 7–8 9–10 11–12 13 14 15 16 17–18 19

Motion switch Surge control (left) Surge control (right) Heave control (front) Heave control (rear) Lamp tune NULL NULL NULL SCU runtime Check sum

ROV Thrusters, sonar and camera etc. Auto-depth, auto-speed etc. Low and high bytes Low and high bytes Low and high bytes Low and high bytes 0 V–5 V Spare Spare Spare Unit: second From byte 3 to byte 18

⎡1 0 ⎢ ⎢0 0 L = ⎢⎢ ⎢ 0 0 Lay ⎢L ⎢⎣ by Lbby 0

⎤ ⎡T0 ⎤ ⎥ ⎢ ⎥ ⎢ ⎥ 1 ⎥ ⎥ T = ⎢T1 ⎥ ⎢T2 ⎥ Lay ⎥⎥ ⎢ ⎥ ⎥ ⎢T ⎥ 0 ⎥⎦ ⎣ 3⎦

(2)

where Li is the moment arms and Tj is the moments. In addition, the vectors of the centre of gravity and the centre of buoyancy of the ROV with coordinates in the body-fixed frame are rG = [xg, yg, zg]T and rB = [xb, yb, zb]T, respectively. From the point of view of equilibrium stability, it is required that xg ≈ xb, yg ≈ yb, zg − zb = h, where h is a stability indicator. The general requirements are not less than 0.01 m in the water surface and not less the 0.02 m in the water. The h of ROV are 0.013 m and 0.024 m, respectively. The architecture of the controller in SMU-II is shown in Fig 10. First, the classic proportional, integral and derivative (PID) approach is adopted for autopilot design, and Ziegler-Nichols tuning rules are applied to produce good values for the PID parameters (Åström et al., 1993; Fossen, 1994; Han and Chung, 2014). According to the structure character and the demand of ROV, a four-degree equation is set up. Based on the motion model of ROV, a fuzzyPID is designed for both depth control and heading control. The fuzzy-PID controller uses the fuzzy rules to online tune the PID parameters, as shown in Fig 10. Here, ud denotes the vector of PID controller outputs, and w denotes the current heading or depth with ordinates in the earth-fixed frame. The initial setting value of heading or depth is r. The output of fuzzy reasoning is a membership function (or fuzzy set), reflecting the fuzzy nature of the control language. The precision process is to find out the exact amount of the fuzzy inference resulting from the fuzzy output membership function. There are a variety of different methods of precision. The centre of gravity method is used to calculate the exact value of the output control. The parameters of the fuzzy query table can be obtained through the operation of the control rules ΔKP, ΔKI and ΔKD . In the actual system design, the fuzzy query table will be generated offline in the system memory. In the control process, the system output control quantisation value is acquired directly (that is, PID controller parameter increment ΔKP , ΔKI and ΔKD) according to the current sampling time of the vehicle and after

03-SUT277-35(1)-ut66627.indd 20

3/28/18 4:25 PM

Underwater Technology Vol. 35, No. 1, 2018

the fuzzy query in memory stored in the fuzzy query table. Then, the control parameters of PID controller can be designed in the following equation: ⎧⎪K P ⎪⎪ ⎪⎨K ⎪⎪ I ⎪⎩⎪K D

K P 0 + a ΔK P K I 0 + b ΔK I K D 0 + c ΔK D

(3)

4. Experiment The tank experiments are carried out to verify the performance of this control system with the various parts of the whole system fully integrated. The water tank is located in Institute of Underwater Engineering, Table 7: Hydrodynamic parameters of the SMU-II ROV

where a, b and c are parameter adjustment factors. These can be adjusted according to the actual control situation, thus changing the parameter increment and the adjustment range. The range of the output variable does not need to be adjusted. In this experiment, it is set as a = 0.3, b = 0.03 and c = 0.01. Fig 11 shows the simulated experimental results for fuzzy PID and PID controller, demonstrating that they can both effectively reach the predetermined value, and the accuracy is essentially consistent. However, in the convergence speed, the fuzzy PID controller is significantly faster, and it is obvious that the fuzzy PID controller is more effective for control system of SMU-II.

Parameter Value

Unit

Brief description

Xu Xu. Xu|u| Zw Zw. Zw|w| Nr Nr|r| xb yb zb Lay

N · s/m N2 · s/m N2 · s2/m2 N · s/m N · s2/m N · s2/m2 N · m · s/rad N · m · s/rad2 m m m m

Linear drag in u Linear inertial drag in u Quadratic drag in u Linear drag in w Linear inertial drag in w Quadratic drag in w Linear drag in r Quadratic drag in r Centre of buoyancy in u Centre of buoyancy in v Centre of buoyancy in w y-distance from the CG to T2 and T3 y-distance from the CG to T0 and T1

–45.4 –62.8 –70.3 –184.6 –476.2 –180.4 –15.73 –4.43 0 0 –0.024 0.18

Lby

0.23 m

(a) (a)

Auto-heading

Auto-heading 220 200

180 Heading (°)

Heading (°)

160 140 120 100

80 60

fuzzy PID PID

0 0

(b)

4 5 6 Time (s)

10 15 20 25 30 35 40 45 50 Time (s)

(b)

Auto-depth

7 2.2 6 2 1.8 4

Depth (m)

1.6 1.4

1.2

fuzzy PID PID

0 0

4 5 6 Time (s)

Fig 11: Simulated experimental results for fuzzy PID and PID controller

15 20 Time (s)

Fig 12: The experiment result of SMU-II in tank test

03-SUT277-35(1)-ut66627.indd 21

3/28/18 4:25 PM

Deng et al. Design and implementation of a remotely operated vehicle testbed

China. The pool is 28 m (length) × 25 m (width) × 5 m (maximum depth). The values of hydrodynamic parameters of the SMU-II ROV are shown in Table 7. To prove the effectiveness of the designed controller, several experiments were carried out in the water tank. Fig 12(a) shows the auto-heading result where the desired heading angle is set to 80°, and the controller shows oscillatory heading error around 2.5°. The auto-depth experimental data are plotted in Fig 12(b), where the target depth is 2.0 m and oscillatory depth error is around 0.2 m. The controller has good application in both depth and direction control, and this is a reasonable amount of error in most engineering applications.

5. Conclusion This paper introduces the design of the SMU-II, and highlights some of the characteristics of engineering applications for ROV, such as continuous working hours, simple crawling function, and perception of the environment in contrast to SMU-I. The paper also describes the implementation of the underwater system and the SCU, and the communication protocol and control design are also given briefly. Compared to the SMU-I, the SMU-II has made significant improvements in drag reduction, endurance capability, and easy manufacture and maintenance. Further improvement of the vehicle has been scheduled: other sensors, such as Doppler velocity logs (DVLs) and ultra-short baseline (USBL) beacons, will be installed to improve its navigation system through underwater tracking control experimental tests. On the basis of this, the trajectory tracking algorithm will be applied to achieve automatic detection of dams and bridges.

Acknowledgment This project is supported by the National Natural Science Foundation of China (51279098, 51409156, 51575336).

References Åström KJ, Hägglund T, Hang CC and Ho WK. (1993). Automatic Tuning and Adaptation for PID Controllers – a Survey. Control Engineering Practice 1: 699–714. Anwar I, Mohsin MO, Iqbal S, Abideen ZU, Rehman AU and Ahmed N. (2016). Design and fabrication of an underwater remotely operated vehicle (Single thruster configuration). 13th International Bhurban Conference on Applied Sciences and Technology (IBCAST). 12–16 January, Islamabad. 547–553. Bandyopadhyay PR. (2005). Trends in biorobotic autonomous undersea vehicles. IEEE Oceanic Engineering 30: 109–139.

Behar AE, Chen DD, Ho C, McBryan E, Walter C, Horen J, Foster S, Foster T, Warren A, Vemprala SH and Crowell JM. (2015). MSLED: The micro subglacial lake exploration device. Underwater Technology 33: 3–7. Chen L, Wang S, Hu H, Gu D and Liao L. (2015). Improving Localization Accuracy for an Underwater Robot with a Slow-Sampling Sonar through Graph Optimization. IEEE Sensors Journal 15: 5024–5035. Deng Z, Zhu D, Xu P, Fang J. (2014). Hybrid underwater vehicle: ARV design and development. Sensors & Transducers 164: 278–287. Fischer N, Hughes D, Walters P, Schwartz EM and Dixon WE. (2014). Nonlinear RISE-Based Control of an Autonomous Underwater Vehicle. IEEE Transactions on Robotics 30: 845–852. Fossen TI. (1994). Guidance and Control of Ocean Vehicles. Chichester: John Wiley & Sons Ltd, 494 pp. Han J and Chung WK. (2014). Active Use of Restoring Moments for Motion Control of an Underwater VehicleManipulator System. IEEE Journal of Oceanic Engineering 39: 100–109. Jiang X, Feng X and Wang D. (2000). Unmanned Underwater Vehicles. Liaoning: Liaoning Science and Technology Publishing House, 464 pp. Javdani S, Fabian M, Carlton JS, Sun T and Grattan KTV. (2016). Underwater Free-Vibration Analysis of Full-Scale Marine Propeller Using a Fiber Bragg Grating-Based Sensor System. IEEE Sensors Journal 16: 946–953. Nadal-Serrano JM and Lopez-Vallejo M. (2015). A Survey on Theoretical and Practical Aspects of Imaging Aids for Artificial Vison in Professional Environments. IEEE Sensors Journal 15: 2719–2731. Page BR, Ziaeefard S, Pinar AJ and Mahmoudian N. (2017). Highly Maneuverable Low-Cost Underwater Glider: Design and Development. IEEE Robotics and Automation Letters 2: 344–349. Ramadass GA, Ramesh S, Subramanian AN, Sathianarayanan D, Ramesh R, Harikrishnan G, Pranesh SB, Dossprakash V and Atmanand MA. (2015). Deep ocean mineral exploration in the Indian Ocean using Remotely Operated Vehicle (ROSUB 6000). 2015 IEEE Underwater Technology Conference (UT), 23–25 May, Chennai, India. 1–8. Salgado-Jimenez T, Gonzalez-Lopez JL, Martinez-Soto LF, Olguin-Lopez E, Resendiz-Gonzalez PA and BandalaSanchez M. (2010). Deep water ROV design for the Mexican oil industry. In: Proceedings of the IEEE OCEANS Conference, 24–27 May Sydney, Australia. 1–6. Shen Y, Hu P, Jin S, Wei Y, Lan R, Zhuang S, Zhu H, Cheng S, Chen J, Wang D and Liu D. (2016). Design of Novel Shaftless Pump-Jet Propulsor for Multi-Purpose Long-Range and High-Speed Autonomous Underwater Vehicle. IEEE Transactions on Magnetics 52: 1–4. Singh Y, Bhattacharyya SK and Idichandy VG. (2017). CFD approach to modelling, hydrodynamic analysis and motion characteristics of a laboratory underwater glider with experimental results. Journal of Ocean Engineering and Science 2: 90–119. Zhang Y and Yu Y. (2011). Core of VxWorks, Device driver and Development detail of BSP. Beijing: Posts & Telecom Press, 323pp. Zain ZM, Harun N, Watanabe K and Nagai I. (2015). Comparison of an X4-AUV performance using a direct Lyapunov – PD controller and backstepping approach. In: Proceedings of the 10 Asian Control Conference (ASCC), 31 May– 3 June, Kota Kinabalu. 1–6.

03-SUT277-35(1)-ut66627.indd 22

3/28/18 4:25 PM

doi:10.3723/ut.35.023 Underwater Technology, Vol. 35, No. 1, pp. 23â&#x20AC;&#x201C;30, 2018

Technical Paper

www.sut.org

De-noising algorithm for SNR improvement of underwater acoustic signals using CWT based on Fourier Transform Kishore Kumar PC*, Sathish Kumar P, Jerritta S and Rajendran V VELS University, Chennai, India Received 10 August 2017; Accepted 17 January 2018

Abstract Underwater wireless communication is a prominent research ďŹ eld owing to its vast range of applications such as underwater sensor networks, remotely operated vehicles (ROVs) and tsunami warning systems. The acoustic signals used in the underwater wireless systems are 105 times slower than light which makes it band-limited. Communication under water is affected not only by natural activities but also by shipping noises, all of which degrade the performance of underwater systems. Hence appropriate de-noising algorithms have to be developed to improve the performance of the band-limited underwater systems. The algorithm described in this paper was developed using linear frequency modulation (FM) waveform as input and using the continuous wavelet transformation (CWT) based on fast Fourier Transform (FFT) method with Morlet wavelet. The proposed algorithm provides an SNR improvement of around 12 dB when compared with the algorithm developed using the chirp signal as input and using Morlet wavelet with wavelet packet decomposition (WPD) technique. Keywords: De-noising, signal-to-noise ratio, SNR, continuous wavelet transformation, CWT, acoustic signal

1. Introduction Underwater wireless communication has gained traction as an important research field owing to the vast range of applications such as underwater seismic monitoring, remotely operated vehicles (ROVs) and disaster warning systems. The underwater systems are band-limited because the acoustic signals used for communication are 105 times slower than light. Communication through sea water is affected by underwater ambient noise (Urick, 1984; Dahl et al., * Corresponding author. Email address: kishore.se@velsuniv.ac.in

2007; Sivakumar and Rajendran, 2010; Sadaf et al., 2015). The ambient noise may be caused by natural activities (e.g. flora and fauna, rainfall and wind) or manmade activities (e.g. fishing boats). These activities are responsible for the degradation of the band-limited underwater systemsâ&#x20AC;&#x2122; performance. Hence algorithms need to be designed to remove ambient noise, thereby increasing the performance of the underwater systems. Techniques like adaptive filters (Yadav and Sharma, 2015), matched phase noise reduction and frame-based time-scale filters (Ou et al., 2011) are used for removal of ambient noise from acoustic signals. However, these techniques are not effective in removing the noise. Prior information of the noise spectrum is needed for the operation of the aforementioned algorithms. Wavelet theory, which provides multi-resolution analysis, overcomes the above drawbacks. The multi-resolution analysis provides a better understanding of the time domain signals. Wavelet theory is based on the principle of autocorrelation, where the wavelet designed similar to the original signal is correlated with the noisy signal. Thresholds are applied to the resulting wavelet coefficients to remove the coefficients pertaining to noises. The original signal is reconstructed by applying the inverse wavelet transform. The algorithm designed with chirp signal as input using Morlet wavelet (Raj and Murali, 2013) based on the wavelet packet decomposition (WPD) (Gokhale and Khanduja, 2010; Chen and Zhang, 2011) is not verty effective and provides a signal-to-noise ratio (SNR) improvement of around 7 dB. The proposed algorithm is developed using the linear frequency modulation (FM) waveform as the 23

04-SUT292-35(1)-ut66628.indd 23

3/27/18 1:48 PM

Kumar et al. De-noising algorithm for SNR improvement of underwater acoustic signals using CWT based on Fourier Transform

input and the continuous wavelet transform (CWT) based on fast Fourier Transform. Here, we explain the proposed algorithm and compare it's performance to that of other proposed solutions.

2. Literature review The technique of memory-less noise suppressing non-linearity in the adaptive filter of an acoustic echo canceller based on normalised least mean square (LMS) was investigated by Wada and Juang (2009). They concluded that both semi-blind source separation (SBSS) based independent component analysis (ICA) and error enhancement procedure are well correlated. These methods help to distinguish the desired signal from other noises or disturbing signals. Jarrot et al. (2005) handled the difficulty of recovering the signal from underwater ambient noise in the environment of multipath propagation of underwater signal. White noise cannot be overlooked because of the distinct spectral characteristics. Hence the classical time estimation method cannot be used here. To overcome this disadvantage, a new method was introduced using unitary warping, which converts a nonlinear signal to a commensurate linear one. The effects of ocean mediation and disruption of underwater background ambient noise were studied by Tu and Jiang (2004). The signal exhibited random process and time-varying characteristics. Hence the de-noising procedure included wavelet transformation of the underwater acoustic signal, and a threshold of the wavelet coefficient and inverse wavelet transformation for reconstructing the received signal. Wang et al. (2009) studied several de-noising techniques, such as wavelet shrinkage threshold (WST), genetic matching pursuit (GAMP) and general matching pursuit (GMP). The WST method held back white noise for low frequency signals, with a narrow frequency bandwidth attributed to the dyadic frequency partition of discrete wavelet transform (DWT). The de-noising process in GMP and GAMP are similar, with time consumption being the only difference. It is essential to maintain the high frequency component for high frequency signals in de-noising the broad bandwidth signals, which was best suited for the GAMP based method. It is theoretically proven that the GMP method had better accuracy in maximum cases. The GMP method, requires very high speed computers with large memories, for complex calculations, which increases the time consumption of GMP. GAMP based de-noising techniques are still difficult in real time for sound signals with high sampling rate.

Hou et al.’s (2008) research on signals with low signal-to-noise ratio shows that the recovery of suppressed signal in ambient noise is very important in the detection and identification process of such signals. They proposed using pangas (a type of fishing boat) for de-noising ship produced noise. In doing so, they achieved a proper estimation of signals using dual-tree CWT. The proposed model was analysed with both ship radiated noise and ambient noise, and proved through experimental measurement. Aggarwal et al. (2011) proposed DWT algorithmbased voice signal de-noising for both hard and soft thresholding. This analysis was performed on voice signal corrupted by babble noise at several SNR levels. The SNR and means square error (MSE) were calculated and compared using two type of thresholding methods. It was observed that hard thresholding is less efficient than soft thresholding.

3. Methodology The proposed method develops a de-noising algorithm using CWT based on FFT (Komorowski and Pietraszek, 2016). The input is the linear FM waveform generated using the ‘phased linear FM waveform’ command in MATLAB. The generated linear FM signal was contaminated using real-time ambient noise data collected from the shallow waters of Bay of Bengal along the Chennai coast. Ambient noise data were removed using the de-noising algorithms developed through CWT and FFT. The effectiveness of the proposed algorithm was evaluated using SNR as the performance metric.

3.1. Collection of ambient noise data Ambient data were collected using a reasonably equipped boat in the shallow waters of Bay of Bengal at a depth of 25 m below sea level. Calibrated omnidirectional hydrophones with a receiving sensitivity of 170 dB and frequency ranging from 0.1 Hz to 25 kHz were used to measure the acoustic pressure of ambient noise. The data collection setup is shown in Fig 1. The data acquisition system (DAS), including computers, hydrophones and a power supply, was secured on a boat. The hydrophones were fixed on a mounted L-shaped setup that was then immersed into the shallow water. The hydrophones, were balanced by weights as shown in Fig 1. The hydrophones measured ambient noise that occurred in the sea coming from various sources, such as fish, sea animals, sea traffic and ships, and the DAS received and stored the data. The specification of the omnidirectional hydrophones made of piezo-resistive material is: 12–24 operating voltage; –2°C to 55°C operating temperature;

04-SUT292-35(1)-ut66628.indd 24

3/27/18 1:48 PM

Underwater Technology Vol. 35, No. 1, 2018

where *

ψ a ,b

* ψ (aw )e jwb

(4)

∞

x (w ) = ∫ x(t )e

(5)

−∞

The coefficients obtained using CWT using FFT were then filtered based on a threshold computed by using the characteristics of the noise data. The signal was then reconstructed. In this work, the derivative of Gaussian (DOG) wavelet and Mortlet wavelet were used to understand the de-noising capabilities. Fig 1: Data collection setup

4. Results and discussion 600 m operating depth; 700 m survival depth; up to 25 kHz operating frequency; and up to –170 dB sensitivity.

The linear FM signal, which is the general characteristic of the pulses transmitted under water, is generated as shown in Fig 2. Fig 3 presents the frequency spectrum of the generated linear FM signal.

3.2. De-noising algorithm The CWT is generally expressed as the correlation of the analysed noisy signal x(t) and the wavelet function ψ(t ), and is defined by the following equation: Cw (a , b )

1 a

∫

∞

−∞ −

⎛t b ⎞⎟ x(t )ψ * ⎜⎜⎜ ⎟ dt ⎝ a ⎟⎠

(1)

where a represents the scaling or dilation, b indicates the time-shifting or translation parameter and Cw (a,b) corresponds to the wavelet coefficients. The CWT requires considerable time for the calculation of the wavelet series. Hence, methods to accelerate the calculation of the CWT efficient algorithms were developed. One such algorithm is using FFT to calculate CWT. Equation 1 can be rewritten in the form of convolution as follows: Cw (a , b )

∫

∞

−∞ −

x(t )ψ *a (b − t )dt

Fig 2: Gated linear FM

(2)

It can be observed that equation 2 represents the CWT obtained by the convolution of the chosen wavelet and the signal to be analysed, which is the linear FM signal contaminated using ambient noise data. CWT can also be expressed in terms of inverse Fourier transform, as shown in equation 3. Equation 4 represents FFT of wavelet function ψ(t ) and equation 5 represents the FFT of the analysed signal x(t). ω indicates the frequency of the signal x(t). This makes CWT a simple convolution of the wavelet and signal at different locations. Cw (a , b )

* 1 ∞ X (w )ψ a ,b (w )dw ∫ −∞ − 2π

(3)

Fig 3: Spectrum of gated linear FM

04-SUT292-35(1)-ut66628.indd 25

3/27/18 1:48 PM

Kumar et al. De-noising algorithm for SNR improvement of underwater acoustic signals using CWT based on Fourier Transform

The linear FM signal transmitted in the ocean is contaminated by the ambient noise caused by multiple artefacts in the ocean. Fig 4 represents the ambient noise collected in the shallow waters of Bay of Bengal. This ambient noise signal was then

added to the linear FM signal, and the resultant signal is as shown in Fig 5. The CWT using FFT was performed on the noisy linear FM signal using Paul, DOG and Mortlet wavelets. Thresholding was performed on the wavelet coefficients after which the inverse wavelet transform was applied to reconstruct the signal. Fig 6 represents the signal decomposition, and Fig 6(b) is similar in frequency and shape to the linear FM signal. Similarly, Figs 7 and 8 indicate the signal decomposition using DOG and Mortlet wavelets, respectively.

Fig 4: Noise signal

Fig 5: The gated linear FM signal with additive ambient noise

Fig 6: (a) CWT performed on the noisy linear FM signal using Paul wavelet; and (b) signal reconstructed using Paul wavelet

04-SUT292-35(1)-ut66628.indd 26

3/27/18 1:48 PM

Underwater Technology Vol. 35, No. 1, 2018

Fig 7: (a) CWT performed on the noisy linear FM signal using DOG wavelet; and (b) signal reconstructed using DOG wavelet

Fig 8: (a) CWT performed on the noisy linear FM signal using Mortlet Wavelet; and (b) signal reconstructed using Mortlet wavelet

04-SUT292-35(1)-ut66628.indd 27

3/27/18 1:48 PM

Kumar et al. De-noising algorithm for SNR improvement of underwater acoustic signals using CWT based on Fourier Transform

The frequency spectrum of the de-noised signal is shown in Fig 9. It can be seen that the spectrum is similar to that of Fig 3, indicating that both the signals are similar in frequency. The SNR was used as the performance metric for evaluating the effectiveness of the proposed algorithm. Previous research using chirp signal as the input and based on Morlet wavelet using WPD and other denoising techniques were able to provide an SNR improvement of 8 dB (Kalpana et al., 2014). Tables 1 to 3 present the input SNR, output SNR and the improvement in SNR obtained at the various scales for Paul, DOG and Mortlet wavelets, respectively. As these tables show and in Fig 10, the Mortlet wavelet has the best improvement in SNR of 12 dB among the three wavelet types. Paul wavelet has an improvement of 5 dB, whereas the improvement in SNR for the DOG wavelet is the least among the three types. Though the Mortlet wavelet provides a better SNR of 12 dB, the number of scales is 18, indicating

the computational complexity and cost is high. Similarly, Paul wavelet also shows improvement on scale 12, which is again computationally high. A trade-off needs to made between the improvement in SNR and computational cost involved.

Fig 9: Frequency spectrum the de-noised linear FM signal

Table 1: SNR calculation using Paul wavelet Scale

Spacing

No. of scales

Input SNR

Output SNR

Improvement in SNR

4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02

15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15

–16.4805 –16.4805 –16.4805 –16.4805 –16.4805 –16.4805 –16.4805 –16.4805 –16.4805 –16.4805 –16.4805 –16.4805 –16.4805 –16.4805 –16.4805 –16.4805 –16.4805 –16.4805

–24.12535 –23.33733 –21.84423 –18.0039 –15.28191 –13.44578 –12.12281 –11.36741 –11.04461 –11.0629 –11.28844 –11.69507 –12.17331 –12.76235 –13.41826 –14.21983 –15.18508 –16.2597

–7.644896684 –6.856876461 –5.36377399 –1.523440317 1.198544562 3.034672202 4.357646403 5.113042 5.435847513 5.417555089 5.192019899 4.785380766 4.30714355 3.718106129 3.062195235 2.260624364 1.295378659 0.220754701

Table 2: SNR calculation using DOG wavelet Scale

Spacing

No. of scales

Input SNR

Output SNR

Improvement in SNR

2 4 6 8 10 12 14 16 18 20

0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02

15 15 15 15 15 15 15 15 15 15

–16.480456 –16.480456 –16.480456 –16.480456 –16.480456 –16.480456 –16.480456 –16.480456 –16.480456 –16.480456

–23.008565 –15.623564 –18.751397 –20.490776 –22.201144 –23.278806 –22.938005 –21.657182 –20.18 –19.097802

–6.528109319 0.856891362 –2.270941745 –4.010320161 –5.720688497 –6.798350106 –6.457549435 –5.176726342 –3.699544254 –2.617346682

04-SUT292-35(1)-ut66628.indd 28

3/27/18 1:48 PM

Underwater Technology Vol. 35, No. 1, 2018

Table 3: SNR calculation using Mortlet wavelet Scale

Spacing

No. of scales

Input SNR

Output SNR

Improvement in SNR

4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02

15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15 15

–16.4805 –16.4805 –16.4805 –16.4805 –16.4805 –16.4805 –16.4805 –16.4805 –16.4805 –16.4805 –16.4805 –16.4805 –16.4805 –16.4805 –16.4805 –16.4805 –16.4805 –16.4805

–22.771723 –24.558999 –27.556428 –22.841202 –19.883483 –18.453052 –20.418402 –19.99408 –17.083497 –14.432071 –10.770859 –7.3956727 –5.4517256 –4.4733611 –4.3765719 –5.0520875 –6.7259066 –9.603629

–6.291267577 –8.078543287 –11.07597261 –6.360746654 –3.403027769 –1.972596841 –3.937946646 –3.513624809 –0.603041036 2.048384375 5.709596165 9.084782793 11.02872998 12.00709443 12.1038836 11.42836807 9.754548934 6.876826504

Fig 10: Comparison of the maximum improvement in SNR of the different wavelets

5. Conclusion This paper shows that the acoustic signals used for underwater communication are affected by fishing activities, flora and fauna, thereby degrading the performance of band-limited underwater systems. The proposed algorithm using the CWT using FFT is analysed using the linear FM signal as input and the real-time noise data collected at Bay of Bengal, Chennai. The simulated results are compared with the algorithm developed using the chirp signal as input with Morlet wavelet, and also with the available wavelets in the CWT using FFT. By comparing the simulation results of the other wavelets with the proposed algorithm, the Morlet wavelet provided a better SNR improvement of 12 dB.

References Aggarwal R, Singh JK, Gupta VK, Rathore S, Tiwari M and Khare A. (2011). Noise reduction of speech signal using wavelet transform with modified universal threshold. International Journal of Computer Applications 20:14–19.

Chen S and Zhang H. (2011). Detection of underwater acoustic signal from ship noise based on WPT method. In: Proceedings of the Fourth International Workshop on Chaos-Fractals Theories and Applications (IWCFTA), 19–22 October, Hangzhou, China. Dahl PH, Miller JH, Cato DH and Andrew RK. (2007). Underwater ambient noise. Acoustics Today: 23–33. Gokhale MY and Khanduja DK. (2010). Time domain signal analysis using wavelet packet decomposition approach. International Journal of Communications, Network and System Sciences 3: 321–329. Jarrot A, Ioana C and Quinquis A. (2005). Denoising underwater signals propagating through multi-path channels. In: Proccedings of Oceans 2005 – Europe, 20–23 June, Brest, France, 501–506. Kalpana G, Rajendran V and Sakthivel Murugan S. (2014). Study of denoising techniques for SNR improvement fort underwater acoustic communication, Journal of Marine Engineering and Technology 13: 29–36. Komorowski D and Pietraszek S. (2016). The use of continuous Wavelet Transform based on the Fast Fourier Transform in the analysis of multi-channel electrogastrography recordings. Journal of Medical Systems 40: 1–10. Kumar PS, Kish Kumar PC, Jerritta S and Rajendran V. (2017). Analysis of wind speed distributions for weekly data. ARPN Journal of Engineering and Applied Sciences 12: 2419–2422. Ou H, Allen JS and Symos VL. (2011). Frame-based timescale filters for underwater acoustic noise reduction. IEEE Journal of Ocean Engineering 36: 285–297. Raj AS and Murali N. (2013). Morlet Wavelet UDWT Denoising and EMD based Bearing Fault Diagnosis. Electronics 17: 1–8. Sadaf S, Yashaswini P, Halagur S, Khan F and Rangaswamy S. (2015). A literature survey on ambient noise analysis for underwater acoustic signals. International Journal for Computer Engineering and Sciences 1: 1–9. Sivakumar VG and Rajendran V. (2010). Analyze the coherence of ambient noise in the bay of Bengal ocean region. In: Proceedings of Recent Advances in Space Technology Services and Climate Change, 13–15 November, Chennai, India, 433–436.

04-SUT292-35(1)-ut66628.indd 29

3/27/18 1:48 PM

Kumar et al. De-noising algorithm for SNR improvement of underwater acoustic signals using CWT based on Fourier Transform

Tieshuang H, Jinglin X, Peng H and Jie S. (2008). Research of s-radiated signal denoising based on the CWT statistical wavelet mode. In: Proceedings of the 9th International Conference on Signal Processing (ICSP), 26–29 October, Beijing, China, 2851–2854. Tu CK and Jiang YY. (2004), Development of noise reduction algorithm for underwater signals. IN: Proceedings of the 2004 International Symposium on Underwater Technology, 20–23 April, Taipei, Taiwan, 175–179. Urick RJ. (1984). Ambient Noise in the Sea. Washington, DC: Undersea Warfare Technology Office, Naval Sea systems Command, Dept. of the Navy, 194 pp. Wada TS and Juang BH. (2009). Acoustic echo cancellation

based on independent component analysis and integrated residual echo enhancement. In: Proceedings of IEEE workshop on applications of Signal Processing to Audio and Acoustics, 18–21 October, NewPaltz, New York. Wang S, He H and Chen XW. (2009). Comparison and application of signal denoising techniques based on time-frequency algorithms. In: Proceedings of the IEEE Intelligent Vehicles Symposium, 3–5 June, Xi’an, China. Yadav R and Sharma D. (2015). De-noising of speech signal using adaptive filter algorithms, International Journal of Technology Enhancements and Emerging Engineering Research 3: 24–28.

04-SUT292-35(1)-ut66628.indd 30

3/27/18 1:48 PM

A Farewell to Ice By Peter Wadhams Published by Allen Lane

Hardcover edition, 2016 ISBN 978-0-241-00941-3 256 pages Described in the front matter as a report from ‘the frontline of planetary change’, A Farewell to Ice presents a powerful call to arms for action on climate change through the lens of its impact on the Arctic. Over 250 pages, Professor Peter Wadhams explores the Arctic’s diminishing sea ice cover and establishes its importance as a driver of global change. Wadhams begins by introducing himself as polar researcher whose career has been devoted to observing and understanding the physical processes that occur in sea ice, largely through extensive expeditions to both polar regions. Over nearly five decades, he has seen first-hand the transition from a white to a blue Arctic, as the snow-covered sea ice has retreated and open ocean is left. In the first two chapters we are introduced to ice, ‘the magic crystal’, in its most basic form and later as the more complex sea ice. As I studied for my own PhD in Arctic sea ice, Wadhams quite literally wrote the textbook that I opened nervously on my first day. Whilst this latest offering is much more than just a textbook, Wadhams still describes sea ice formation and behaviour, as well as its importance for global climate, with the same clarity that I found so accessible back then.

The descriptions of even the most basic chemistry and physics of ice are almost poetic, and his passion for the subject makes much of this book a delight to read. The next three chapters take us through a history of the Earth’s climate over the past hundreds of millions of years to present. As one would expect, much of the discussion focuses on ice and the ice ages. Wadhams describes the periodic astronomical forcing of ice ages. He moves on to explain how human-driven increases in atmospheric greenhouse gas concentrations may impact on this cycle, raising the alarming question: have we created our own geological age? We are then introduced to the concept of ‘Arctic amplification’ – where Arctic warming mimics global warming through the past century but in an amplified fashion. In chapters 6 to 9, we reach the heart of the book. If the previous chapters had been a warm up, then here Wadhams ups the ante to rally the troops to his crusade against climate change inaction. We are presented with the first detailed description of his own work, primarily in sea ice thickness measurements from submarine sonar. In 1990, he provided the first evidence that Arctic sea ice was thinning, in addition to the decrease in extent that had been measured by satellites since 1979. These field measurements were crucial because no satellite could measure sea ice thickness at that time. Chapter 7 describes the ‘Arctic death spiral’ – the idea that the decline in Arctic sea ice volume is accelerating towards an imminent disappearance in summer, and an inevitable tipping point

www.sut.org

Book Review

doi:10.3723/ut.35.031 Underwater Technology, Vol. 35, No. 1, pp. 31–32, 2018

from where it cannot recover. The climate feedback mechanisms associated with sea ice decline are covered coherently in chapter 8. Chapter 9 is dedicated solely to Arctic methane and what Wadhams sees as the impending catastrophe for our present-day climate and economy. The hypothesis is an increase in global temperature and resulting sea ice decline will allow the release of methane from offshore permafrost into the atmosphere, which will further warm our climate. It is in these later chapters that the book narrows in focus and slows. Readers would benefit from a measured overview of the different scientific approaches to estimating when the Arctic will become ‘sea ice free’ (defined in the book as when sea ice extent dips below ~1 million square kilometres), and what drives the variation in these estimates. The same is true for differing scientific opinion regarding the potential magnitude of methane release from the Arctic seabed. Throughout the book, Wadhams repeats claims that the Arctic could be sea ice free by the time of publication in the summer of 2016; we now know that whilst 2016 hailed the second lowest ice extent on record, it was still 3 million square kilometres above ‘ice free’. The summer of 2017 saw the seventh lowest sea ice extent on record. Whilst the longterm climate signal of decreasing sea ice is clear, the catastrophism here feels out of step with other community research. In chapters 10 and 11 balance is restored. Wadhams gives an excellent summary of the potential impact of sea ice loss on global atmospheric and oceanic

05-SUT296-35(1)-ut66629.indd 31

3/27/18 1:48 PM

Wadhams. A Farewell to Ice

circulation, as well as its effect on associated weather, food and water supplies, and agriculture. Chapter 12 briefly transports the reader to Antarctica, asking, ‘Does it matter what happens down there?’ Its answer is that it certainly does. The closing two chapters focus on the need for a global solution to the increase in atmospheric greenhouse gas concentrations. Wadhams warns against denial and inaction, and suggests personal actions that could make a difference (e.g. speak out, reduce energy use, lobby government).

Throughout A Farewell to Ice, Wadhams describes the processes by which sea ice forms, melts and influences our climate with a simplicity and emotion that can only be attained through a lifetime of study. These are made all the more enjoyable when interspersed with personal anecdotes from the field. However, as we enter a new era of sea ice research, I felt Wadhams was too dismissive of the importance – and indeed success – of more modern techniques of exploration, such as satellite observations of

sea ice thickness and climate model predictions. Combining the methods Wadhams pioneered with these new techniques will equip us more readily to understand the mechanisms and implications of sea ice retreat and the time we have left before, as Wadhams himself so eloquently states, “It will not just be a farewell to ice, but a farewell to life”. (Reviewed by Dr Rachel Tilling, Centre for Polar Observation and Modelling, Institute for Climate and Atmospheric Science, University of Leeds)

05-SUT296-35(1)-ut66629.indd 32

3/27/18 1:48 PM

Society for Underwater Technology

Educational Support Fund Sponsorship for Gifted Students in Marine Science, Technology and Engineering to meet industry’s critical shortage of suitably qualiﬁed entrants.

SUT sponsors UK and overseas students (studying in the UK and abroad) at undergraduate and MSc level who have an interest in marine science, technology and engineering. Students are supported who are studying subjects such as:

Offshore and Ocean Technology Subsea Engineering Oceanography Marine Biology Ship Science and Naval Architecture Meteorology and Oceanography The SUT annual awards are £2,000 per annum for an undergraduate, and £4,000 for a one-year postgraduate course. (Part-time postgraduate studies funding available.) As one of the largest non-governmental sources of sponsorship, the SUT has donated grants totaling almost half a million pounds to over 270 students since the launch of the fund in 1990.

For further information please contact Society for Underwater Technology, Unit LG7, 1 Quality Court, London WC2A 1HR UK t +44 (0)20 3440 5535 e info@sut.org or please visit our website

www.sut.org 06-SUT-35(1)-IBC.indd 1

3/27/18 1:58 PM

Headline Sponsor

In association with:

Visit the UKâ&#x20AC;&#x2122;s leading renewable and low carbon energy event JOIN the largest FREE-TO-ATTEND meet up of energy generation solutions and low carbon technology

Learned Society Patron:

Key International Trade Partner:

Host city:

EXPLORE new solutions, network and stay UP-TO-DATE with sector trends

7,500+ ATTENDEES

300+ EXHIBITORS

400+ SPEAKERS POWER TRANSPORT HEAT EFFICIENCY & MUCH MORE. Organised by:

07-SUT-35(1)-OBC.indd 1

all-energy.co.uk/register18 @AllEnergy +44 (0)20 8439 5560

3/27/18 1:58 PM