A Fatigue Measurement and Analysis Method based on GYRO and SEMG for Upper Limb Training of Stroke

www.seipub.org/gpai

Global Perspectives on Artificial Intelligence (GPAI) Volume 3, 2015 doi: 10.14355/gpai.2015.03.003

A Fatigue Measurement and Analysis Method based on GYRO and SEMG for Upper Limb Training of Stroke Patients Naigong Yu1, Chunmin Zhao*1, Hsiao‐Lung Chan2 *1

College of Electronic Information & Control Engineering, Beijing University of Technology, Beijing 100124, China

Department of Electrical Engineering, Chang Gung University, Taoyuan333, Taiwan

2

1yunaigong@bjut.edu.cn ; *1zcmbuaa2009@163.com ; 2chanhl@mail.cgu.edu.tw

Abstract A fatigue measurement and analysis method was developed based on gyroscope kinematics signals and SEMG during the upper limb training of stroke patients. Stroke is the major cause of the elderly upper limb Movement disorders. On clinical, to restore the patientʹs motor function, the rehabilitation way like machine auxiliary training is taken. It has great improvement for the recovery of stroke patients. But fatigue caused by excessive movement training will hurt for stroke patients again. The fatigue detection can reduce damage brought by fatigue. Lack of some physiological signals to provide qualitative assessment of the amount of data, the former fatigue evaluation standard may judge by subjective rehabilitation division. This study combines the SEMG signal and the kinematics parameter collected to make qualitative analysis during the upper limb function rehabilitation training. The gyroscope kinematics signals measure was divided into the corresponding cycle segmentation in accordance with the cycling action. The SEMG signal measured will be made some signal processing analysis in the corresponding action cycle. The change of frequency spectrum analysis of the SEMG signal through the period of time detect the generation of fatigue.

Keywords Stroke Rehabilitation; Robot‐Assist Training; SEMG; Gyroscope; Median Frequency; Fatigue Measurement

Introduction Some research has indicated that fatigue and decrease in muscle strength may serve as important indicators in frailty [1].In fact, both factors are also considered as typical symptoms associated with patients with central nervous system damage [2 – 6]. Movement function in these patients with clinically can be impaired by rehabilitation. Currently no clinical evaluation of qualitative data, only through rehabilitation teacherʹs subjective judgment but the rehabilitation training of excessive will no doubt bring patients of secondary damage. So the patientʹs muscle power information and kinematics characteristics to the evaluation of patients in clinical index in the process of rehabilitation and rehabilitation strategies have important guiding significance.

In general, machine auxiliary training based rhythmic contractions are similar to the control of wrist action since completion of these actions requires agonistic and antagonistic muscles to be alternately activated in coordination with a time sequence. For stroke patients, an auxiliary training machine can provide more adequate support for the wrist and help patients stretch their muscles affiliated to wrist joint. In addition, some researchers have also indicated that such movement is safe, effective and accessible to patients with a wide range of motor impairments [7–9]. Therefore, machine‐based movement is often used for limb training[10]. Previous research in literature even indicates that patients with stroke and cerebral palsy may improve their motor and balance abilities after an early short duration of machine training [11, 12].

Physical fatigue is naturally accompanied by a progressive decline in motor function during training. In this way, a previous research report has proposed an experimental system to study physical fatigue during sustained and dynamic contractions [13]. On the other aspect, it is indicated from another previous study that a prolonged auxiliary exercise would cause decrements in whole‐body power, muscle‐function, and jump‐performance

14

Global Perspectives on Artificial Intelligence (GPAI) Volume 3, 2015 www.seipub.org/gpai

measures, implying that the exhaustion due to cycling exercise are related to fatigue in daily function [14].

Moreover, investigations into alternative measures for muscle fatigue as well as muscle activation during cycling were also conducted in previous research [15–17]. We may thus infer that a system designed based on long‐term repetitive exercise, such as the machine‐based training, may be used as a pattern for continuously, fully tracking and detecting fatigue or degeneration of motor functional activities. Therefore, the aim of this research is to develop a system for producing and quantitatively analyzing the gyroscope kinematics signals data and SEMG signal data to detect the fatigue during training movement.

In this paper, a new robot‐assist training system for monitoring and analysis of fatigue‐related parameters is introduced. In order to achieve the goal of muscle fatigue assessment, this study combines the SEMG signal and the kinematics parameter collected to make qualitative analysis during the upper limb function rehabilitation training.

The Inertial Unit Signal Collection and Processing System In this study, the fatigue detection method is put forward, which is based on the SEMG signal in the process of patient rehabilitation training and the kinematics parameters. In addition to considering the device itself which is easy to carry with proper size and used with low power consumption, the premise of hinder from the patients during the action of rehabilitation training also can be excluded in the process of rehabilitation training.Therefore, we adopted the Texas instruments MSP430F5438 chip microprocessor.

SEMG signal acquisition circuit design system contains the sensor chip components ADS1294 of Texas instruments company, different from the traditional circuit component with bigger volume, higher consumer. Two electrodes should be fixed on the patientʹs muscle movement surface as the SEMG signal collection patch. When the body produces movement, an poor relative electrical potential difference through two electrodes generates physiological signal.After a front‐end amplifier amplification , a set of second order high pass filter and low pass filter, the signal outside the scope of the selected band will be filtered. Disposed by digital and analogy conversion circuit, numerical SEMG signal we need was obtained.

SEMG signal disposed by analogy and digital conversion was finally transmitted through the SPI communication interface and microprocessor , then was connected to the Bluetooth module to Bluetooth part by MSP430 UART interface at the receiving end.The receiver is adopted computer programming with c # program to accept and store data.The whole system is shown in Fig. 1

FIG.1 SEMG SIGNAL COLLECTION SYSTEM

The Kinematics Data Acquisition Process Bi‐Manu‐Track is used as the upper arm auxiliary rehabilitation training machine device. Machine itself has three models: 1. The passive mode: hands were drived to act by the machine itself set in rotational speed and scope according to the status of the subjects .2. Passive active: one hand actively drove another hand .3. Hands are

15

www.seipub.org/gpai

Global Perspectives on Artificial Intelligence (GPAI) Volume 3, 2015

passive: when achieved their respective scope of stress, two hands can act independently.

In this study, SEMG signal acquired when subjects were rained on three training mode of the auxiliary training machine device named as Bi‐Manu‐Track. At the same time, related kinematics parameters can be detected on our subjects.

Small drift is influenced by environmental temperature with high precision. Thus three axis gyroscope sensor were adapted to measure the human body movement kinematics parameters as the signal measuring device. We adopt three axis gyroscope sensor named as LYPR540AH produced by STMicroelectronics company to achieve the periodic motion segmentation gyroscope signal and then choose zero‐crossing way to measure angular velocity in our system .

Principle of Fatigue Detection Based on GYRO movement cycle signal, the SEMG signal will be cut to some corresponding action cycle.

Through the change of frequency during the period of every cycle, the frequency value of the SEMG signals spectrum analysis detect the generation of fatigue.

Fig. 2 TWO CHANNEL SEMG AND GYRO SIGNALS

Methods and Materials Experimental Procedure Before rehabilitation training, there basically was both actions. These were the Extension and Flexion channel, the pronation and supination channel. Position measurement of the pronation and supination channel muscle group is relatively deep, not easily detected. The tests should be familiar with the relevant actions, and then do some pattern actions under controlled formal action patterns. The extensor and flexor channel was chosen to represent the SEMG signal. After sampling ADC values, the signal should be converted to the voltage value. SEMG signal is concentrated in the 50‐150 Hz and filtered. Three‐axis gyroscope is mainly concentrated between 0.7 1 Hz.

16

Global Perspectives on Artificial Intelligence (GPAI) Volume 3, 2015 www.seipub.org/gpai

Kinesiological Data Analysis SEMG signal, tri‐axial acceleration and gyroscope signal sampling frequency is 1000 Hz. SEMG signal filter range is 10‐500 Hz, and interference signal of 60 Hz could be filtered out. To measure the SEMG signal, SEMG signal need to be converted to the standardized conversion voltage value after ADC sampling of numerical value.

General SEMG signal mainly is in the 50 ~ 150 Hz band. The process is shown in figure 2.

An Innovative New‐Defined Indicator of Muscle Fatigue—FPM To measure gyroscope signal after filtering correction filtering, cycle Angle change through actions, we can zero ‐ crossing ways to stretch gyroscope signal in each cycle cutting.

It is different from the usual detection methods, interference by gravity. After segmentation of gyroscope, signal can be anyway in the corresponding period extensor and flexor muscle movement, so we analyse corresponding electromyography signal frequency domain analysis of time to achieve the purpose of detection of muscle fatigue.

The Results and Discussion In order to evaluate the performance of the proposed system in actual practice, we built up the system and then used it to collect the raw EMG data derived from VL and GAS muscles of the five healthy subjects during the bipedal cycling movement.

Kinesiological analysis such as FFT‐based PSD, EA and MF calculations were all synchronously performed and displayed by the system software.

In addition, FPM tracings were also generated. All these results were jointly applied for on‐line fatigue assessment.

It should be noted that among all these EMG‐related factors, FPM method is an innovative and new technology. In fact, it can not only characterize the progression of fatigue, but also determine the explicit onset time of the fatigue occurrence, thus representing the most significant benefit from our research. The result is shown in Fig. 2 Therefore, the subsequent outcome analysis and discussion will be more focused on the performance evaluation of FPM method. We can speculate fatigue according to the SEMG signal changes Conclusions The purpose of this thesis study mainly for stroke patients in rehabilitation training when an analysis is made of bio‐medical signal processing, and the results of kinematics of inertia parameters, to evaluate the process of fatigue testing, is to establish related indicators and analysis methods will be able to help more people to know their situation, in the process of rehabilitation science rehabilitation reduce unnecessary muscle damage.

REFERENCES

[1]

Faulkner, J.A.; Brooks, S.V.; Zerba, E. Skeletal muscle weakness and fatigue in old age: Underlying mechanisms. Annu.

Rev. Gerontol. Geriatr. 1990, 10, 147–166.

[2]

Amato, M.P.; Ponziani, G.; Rossi, F.; Liedl, C.L.; Stefanile, C.; Rossi, L. Quality of life in multiple sclerosis: The impact of depression, fatigue and disability. Mult. Scler. 2001, 7, 340–344.

17

www.seipub.org/gpai

[3]

Global Perspectives on Artificial Intelligence (GPAI) Volume 3, 2015

De Groot, M.H.; Phillips, S.J.; Eskes, G.A. Fatigue associated with stroke and other neurologic conditions: Implications for stroke rehabilitation. Arch. Phys. Med. Rehabil. 2003, 84, 1714–1720.

[4]

Merkelbach, S.; Sittinger, H.; Koenig, J. Is there a differential impact of fatigue and physical disability on quality of life in multiple sclerosis? J. Nerv. Ment. Dis. 2002, 190, 388–393.

[5]

Morley, W.; Jackson, K.; Mead, G.E. Post‐stroke fatigue: an important yet neglected symptom. Age Ageing 2005, 34, 313.

[6]

Van der Werf, S.P.; Jongen, P.J.; Lycklama à Nijeholt, G.J.; Barkhof, F.; Hommes, O.R.; Bleijenberg, G. Fatigue in multiple sclerosis: Interrelations between fatigue complaints, cerebral MRI abnormalities and neurological disability. J. Neurol. Sci. 1998, 160, 164–170. with Parkinson’s disease. Acta. Neurol. Scand. 2003, 107, 1–6.

[7]

Chen, H.Y.; Chen, S.C.; Chen, J.J.; Fu, L.L.; Wang, Y.L. Kinesiological and kinematical analysis for stroke subjects with asymmetrical cycling movement patterns. J. Electromyogr. Kinesiol. 2005, 15, 587–595.

[8]

Brown, D.A.; Kautz, S.A.; Dairaghi, C.A. Muscle activity adapts to anti‐gravity posture during pedaling in persons with poststroke hemiplegia. Brain 1997, 120, 825–837.

[9]

Rosecrance, J.C.; Giuliani, C.A. Kinematic analysis of lower‐limb movement during ergometer pedaling in hemiplegic and nonhemiplegic subjects. Phys. Ther. 1991, 71, 334–343.

[10] Brown, D.A.; Nagpal, S.; Chi, S. Limb‐loaded cycling program for locomotor intervention following stroke. Phys. Ther.

2005, 85, 159–168.

[11] Katz‐Leurer, M.; Sender, I.; Keren, O.; Dvir, Z. The influence of early cycling training on balance in stroke patients at the subacute stage. Results of a preliminary trial. Clin. Rehabil. 2006, 20, 398–405.

[12] Williams, H.; Pountney, T. Effects of a static bicycling programme on the functional ability of young people with cerebral palsy who are non‐ambulant. Dev. Med. Child. Neurol. 2007, 49, 522–527.

[13] Kiryu, T.; Morishiata, M.; Yamada, H.; Okada, M. A muscular fatigue index based on the relationships between superimposed M wave and preceding background activity. IEEE Trans. Biomed. Eng. 1998, 45, 1194–1204.

[14] Mcintyre, J.P.; Mawston, G.A.; Cairns, S.P. Changes of whole‐body power, muscle function, and jump performance with prolonged cycling to exhaustion. Int. J. Sports Physiol. Perform. 2012, 7, 332–339.

[15] Balasubramanian, V.; Jagannath, M.; Adalarasu, K. Muscle fatigue based evaluation of bicycle design. Appl. Ergon. 2014, 45, 339–345 [16] Elder, C.P.; Cook, R.N.; Wilkens, K.L.; Chance, M.A.; Sanchez, O.A.; Damon, B.M. A methodfor detecting the temporal sequence of muscle activation during cycling using MRI. J. Appl.Physiol. 2011, 110, 826–833.

[17] Fonda, B.; Panjan, A.; Markovic, G.; Sarabon, N. Adjusted saddle position counteracts themodified muscle activation patterns during uphill cycling. J. Electromyogr. Kinesiol. 2011, 21,854–860.

18