Effect of initial spring compression of four-bar knee joint on EMG signals in AK amputees
- (0) Download
https://www.eduzhai.net American Journal of Biomedical En gineer in g 2013, 3(5): 99-106 DOI: 10.5923/j.ajbe.20130305.01 Effect of Initial Spring Compression in a Four Bar Knee Joint on EMG Signals for an AK Amputee Some r M. D. Nacy1,*, Shake r S. Hassan2, Mohannad Y. H. Al Maqdasi2 1Al Khwarizmi College of Engineering, University of Baghdad, Baghdad, Iraq 2M achines and Equipments Engineering, University of Technology, Baghdad, Iraq Abstract A mod ified four bar knee mechanis m has been designed to study the behavior of EM G act ivity of AK amputee person by the variation of the initial co mpression settings of the assistant spring , fro m -2 mm to +6 mm as co mpared to the initial standard spring setting, and for six variable walking speeds, namely, 0.4, 0.7, 1.1, 1.4, 1.8, 2.0 km/hr. The results show a clear relat ionship between the spring settings and the minimu m EM G activ ity, wh ich permits the future design of an actuator to vary the initia l spring setting for each wa lking speed to reduce the load e xerted by the amputee person. Keywords Above Knee Amputee, Four Bar Knee Mechanism, Electro myography 1. Introduction The four bar knee mechanis m is widely used for A KA lower limb prosthesis because of its simplicity, easy maintenance and low cost. Studies show that, above knee amputees as compared to non amputee individuals, have up to 60% higher energy consumption and utilize three times the power and torque in the h ip. Efforts are continuously being made to develop new knee mechanisms to alleviate amputee discomfo rt, lower energy consumption and instantaneous muscular effort, reduce loads transferred to the sound leg and to the vertebral column, and imp rove stability and appearance in walking. Following the loss of the physiological knee, the hip muscles must control the new knee as well as continuing to control thigh-pelvis angular motion. During stance phase the amputated side of the patient requires increased stability, so flexion of the knee should be prevented. Knee-stability can be gained partially through the align ment of the prosthesis and by the choice of the knee-unit. The design of one category of art ificial knees is based exclusively on mechanisms with mechanical properties, such as friction, spring and damping coefficients, hydraulic and pneumatic valves, which remain constant during the walking cycle o r d epend on knee flexio n-extens ion kinemat ics. Sing le axis AKA prostheses with mechanicalconstant-frict i on knee units are simple, lightweight, dependable, reliab le and commercial availability at reasonable cost but they can * Corresponding author: nacys2@asm e.org (Somer M. D. Nacy) Published online at https://www.eduzhai.net Copyright © 2013 Scientific & Academic Publishing. All Rights Reserved only be set up to walk optimally at one speed. A mo re advanced swing control for prosthetic knees uses fluid dynamics to provide variab le resistance, containing air (pneumatic) or fluid (hydraulic). Although hydraulic knees provide a s moother gait, in co mparison with other knee systems, hydraulic un its are heavier, require mo re maintenance, and cost more. The polycentric, four-bar lin kage prosthetic knee joint has a moving instantaneous center of rotation which lies far pro ximal and posterior to the anatomic knee center, maintaining it posterior to the weight-bearing axis of the lower limb . Th is provides an extension mo ment at the knee, decreasing the potential for knee buckling, and decreasing the risk of the patient falling. Control is obtained using constant friction or a hydraulic damping, and by storing energy in a spring during knee flexion and releasing it during extension. Many research works were achieved to enhance knee performance through damper valves and pneumatic system improvements, and through designing and inserting special purpose control units,[1, 2, 3, 4 and 5]. Since the most useful technique applied to evaluate and record the activity of muscles is electromyography (EM G), in wh ich an electrical potential emanates fro m the muscle itself during its action, hence researchers have adopted this method to measure muscles activity and to indicate the muscles which are most responsible for motion for AK amputee,[6, 7, 8, 9 and 10]. In this work a new low cost method is introduced by modifying a widely used four bar knee mechanis m, wh ich is Otto Bock 3R20, in order to minimize the load consumed by the amputee person while walking at different speeds. There are a number of d ifferent parameters that can be altered during the walking of the above knee amputee in order to obtain better walking characteristics which are mo re similar to the walking of sound person. One of these parameters that 100 Somer M . D. Nacy et al.: Effect of Initial Spring Compression in a Four Bar Knee Joint on EM G Signals for an AK Amputee have been considered in this study is the initial co mpression setting of the assistant spring adjustment. 2. Methodology The present four bar knee mechanism have four control parameters that can be adjusted at different setting for each above knee amputee person in order to achieve a good motion characteristics and best comfort ability for the amputee person, these parameters are: adjusting stance phase stability, axis friction adjustment, prosthesis alignment and initial co mpression setting of the assistant spring adjustment. The first three parameters are fixed. The main parameter taken into consideration is the initial co mpression setting of the assistant spring by adjusting the compression of the spring through turning the adjustment screw of the spring; the maximu m depth of the adjustment screw in the spring housing should not exceed (8 mm). To achieve this task, a special purpose mechanism was designed and imp lemented, as depicted in Figure 1. The lower limb pylon is slotted in the vertical d irection and a vertical slotted threaded cylinder is mounted in transitional fit over the pylon at the same slotted area of the pylon and fixed by four (M 4) threaded bolts. The spring housing and the adjustment screw are removed to permit the extension of the initial setting of the assistant spring, wh ile the knee joint is assembled and aligned with all co mponents of the above knee lower limb assembly. These are replaced by a new larger and longer cy linder have the same upper thread of spring housing but with two successive inner holes, the first one is for housing the spring with the same inner dia meter of the spring housing, the second one is with larger inner diameter to permit a vertical motion of a piston which is used to control the compression of the spring. The lo wer part of the second hole is threaded with a holed nut. The p iston has (M4) inner thread. A p iston rod is threaded with spring piston and moved freely in vertical d irection through the holed nut. The piston rod is threaded at its lower part and assembled with the assembly cylinder wh ich is horizontally threaded to fix the horizontal threaded shaft. The horizontally threaded shaft is mounted through the slotted cylinder of the pylon and vertically adjusted by the outer nut. If the outer threaded nut moves vertically, the init ial compression of the assistant spring force is increased and vice versa. The new housing of the spring can be adjusted at any setting points, from -2 mm fro m the orig inal init ial spring setting to +6 mm fro m the original init ial spring setting. Each one co mplete cycle of the threaded outer nut makes 1 mm extra co mpression of the init ial assistant spring setting. The new modified knee jo int is tested with above knee amputee walking on treadmill at six d ifferent walking speeds, namely, 0.4, 0.7, 1.1, 1.4, 1.8, 2.0 km/h, at nine setting points of the in itial assistant spring setting, namely, -2, -1, 0, +1, +2, +3, +4, +5, +6 mm, for (30 sec.) fo r each speed, hence each 30 sec contains 18-24 gait cycles depending on the walking speed. Each of the above tests were carried for eight times to insure the accuracy of the results. 3. Kinematics of Four Link Mechanism The four bar lin k mechanism of the knee is shown schematically in Figure 2. It is required to obtain the deflection imposed on the assistant spring; hence the force induced on the spring can be calculated. To achieve this task, the following procedure is adopted. By knowing the knee angle (Ɵ knee) as measured by Dartfish software, hence, Ɵk3 = 27o + Ɵ knee (1) Using the trigonometric identities of the four bar mechanis m, hence, Ɵk2 =2*tan-1[−E ± (E2−4*D*F) 1/2/2*D] (2) Where: D = K +a*b− (a*c + b*c)*cos (Ɵk3) (3) E = 2*b*c*sin (Ɵk3) (4) F = K−a*b+ (b*c−a*c)*cos (Ɵk3) (5) K = (a2+b2+c2−d2)/2 (6) Obtaining Ɵk2, hence Ɵs can be calculated as, Ɵs = β – Ɵk2 (7) Where: β = constant = 108o Fro m Ɵs, the deflection of the spring is, DEF = e*sin (Ɵs) (8) Where, e = constant = 15 mm Finally, the force on the spring, obtained from its calibrat ion chart is, Fs = 6.26566 * DEF (9) American Journal of Biomedical En gineer in g 2013, 3(5): 99-106 101 Knee joint assembler Vertical threaded cylinder Vertical slot Outer threaded nut Pylon Four bar knee joint Assistant spring Housing cylinder Holed nut Piston rod Ho rizont al threaded shaft Assembly cylinder Figure 1. The new modified extension assistant spring housing and the external control mechanism of the initial extension assistant spring setting Figure 2. Four bar mechanism, linkages and angles 102 Somer M . D. Nacy et al.: Effect of Initial Spring Compression in a Four Bar Knee Joint on EM G Signals for an AK Amputee 4. Experimentations The AKA person is a 45 years old male, 95 kg weight, 1.76 m height, has a proximal thigh stump in his right limb since 1987 as an accident result. He is an active person, working all the day time wearing the present Otto Bock (3R20) four bar knee mechanism since 2005, with single axis ankle foot assembly, the foot size is 26 with full pylon length. The treadmill used is of type Zebris, FDM -T system for stance and gait analysis affording different walking speeds. The prosthetic lower limb of the amputee person is marked at the hip joint, knee jo int, ankle joint and anterior and posterior points of foot with orange lighted markers to make fu ll indicat ion of the lower limb joint angles during complete gait cycle which are the hip angle, the knee angle and the ankle angle. A camera of type (SONY HX-1 fu ll HD-1080) is used to capture the mobility of the prosthetic during the tested time for each test. Figure 3. Sample EMG test (walking speed = 1.1 km/hr, initial setting of assistant spring = + 2 mm) American Journal of Biomedical En gineer in g 2013, 3(5): 99-106 103 The tested muscles are two muscles for the amputated limb and the other two for the intact limb; these are the rectus femoris (RF) muscle wh ich represents the sample of muscles responsible of flexing the h ip jo int and the gluteus maximus (GM) muscle which represents the sample of muscles responsible of extending the h ip jo int. The electro myograph ic activit ies (EM G) fro m four muscles are collected using myoresearch-xp clinical edit ion1.07.01 standard EM G analysis protocols – NORAXON-USA. By using surface EM G electrodes of type Noraxon disposable, self adhesive AG/A GCL snap of dual electrodes (4 mm) electrode with (20 mm) electrodes spacing, used for surface EM G applications only. Sample EM G chart is shown in Figure 3. The EM G tested electrodes locations which are the (RF) and (GM) area are shaved and cleaned with alcohol and dried with cotton prior to placing the electrodes. In addition to that the electrodes are fixed with adhesive tape to prevent electrode motion fro m the skin location during the tested time. A ground electrode was placed near sacrum area. The muscles electrodes are placed on each muscle in the picture of electrode locations presented by the Noraxon EM G software. Firstly the initial setting of the assistant spring is located at (-2 mm) fro m the Init ial standard manufacturer spring setting and the amputee walked on treadmill at the six selected walking speeds. The selected walking speeds are chosen to represent the range of amputee walking fro m lo wer speed to the upper possible speed that the amputee can walked without problems. Each speed is tested for (30 seconds), the number of gait cycles for the selected time for each speed is different and related to the time of the gait cycle for each tested speed. The above procedure is repeated for other spring settings, namely, -1, 0, +1, +2, +3, +4, +5, +6 mm. The captured video during each test and with the aid of the orange lighted markers and the Dartfish software is used to obtain the behavior of the hip, knee, and ankle jo int angles during a comp lete gait cycle. Using the four bar mechanism kinematics, and knowing the variation of the knee joint angle during a co mp lete gait cycle, the force and deflection induced in the assistant spring can be obtained as variables during the complete gait cycle. Each one second walking time is div ided into twenty points; each point represents 5% of the second (i.e. one frame). The time of each gait cycle (t) can be easily estimated fro m the duration from heel strike to toe off, and therefore the accumulated potential of the spring (AP) for each gait cycle can be calculated using the formula. AP = (Σ0.5*Fs*DEF)/ (0.05*t) (10) The AP represents the applied load the amputee person exerted in each gait cycle. Increase of AP represents an increase in the load exerted by the amputee person and vice v ers a. EM G activ ity signals fro m the selected four hip muscles are measured for thirty seconds and sampled at 1000 Hz sampling rate and stored to the hard disk, the default signal processing consists of full wave rectificat ion of all checked channels and smoothing using root mean square algorithm with a 100 ms time constant. The tested time is divided into a number of periods each period represent the beginning and ending marker points of each muscle activation, the number of periods depend on the speed of the amputee at each test, the marker points of each muscle activation are v isually inspected. For each period the mean amplitude of the muscle activity is calculated, and then the mean of mean amp litudes of all periods is calculated. Each test is rep resented as EM G activity by the mean of mean a mplitude of all periods. 5. Results and Discussions The variation of force on the assistant spring during a complete gait cycle for d ifferent walking speeds at the standard initial assistant spring setting, i.e. at 0 mm init ial compression, is shown in Figure 4. The constant value of the spring force (70.65 N) fro m heel strike (0% gait cycle) to the knee flexion (40-55% of gait cycle) and the same fo r the late of swing phase (85-100% of gait cycle) represents the constant initial co mpression of the spring during the knee locking (knee at full extended position). The higher force fro m (40-55%) to (85-95%) o f gait cycle represents the knee flexion at different walking speeds, late beginning o f knee flexion at low walking speed (0.4 km/h) to early beginning at high walking speed(2.0 km/h) ,with late ending of knee flexion at the slow walking speed and early ending at high walking speed. The variation of the accu mulated potential of the assistant spring with the walking speeds for the (standard, +1mm, +2mm and +3 mm) in itial co mpression of the assistant spring setting is shown in Figure 5. The min imu m accu mulated potential at each gait cycle represents the minimu m load e xerted by the amputee person on the four bar knee assembly during the gait cycle at each walking speed. Energy is saved during knee flexion at the late part of stance phase and released during the knee extension at the late part of swing phase. The speed of knee return fro m the flexing position to the extending position depends on the accumulated deflection of the spring and on the walking speed, the time of spring deflection and spring releasing are responsible for counting the duration of the swing phase and also the walking characteristics of the amputated and intact limbs, therefore the minimu m accu mu lated potential of spring is not the same for all walking speeds. The minimu m accu mulated potential at (0.4, 0.7, 1.1, 1.4, 1.8 and 2 km/h) is found to be at the following initial spring settings (+ 2mm, standard, standard, +2mm, +2mm, +2mm) respectively. The variation of the mean EM G activ ity at the nine tested initial assistant spring settings and for both the right (amputated) limb and the left (intact) limb of the amputee person at the tested walking speeds, is shown in Figures 6 and 7 respectively. Since the EM G results are of statistical nature, and to insure a good scatter of the data, hence the standard deviation was calculated. A samp le o f the calculated standard deviation, for the case of + 2 mm as 104 Somer M . D. Nacy et al.: Effect of Initial Spring Compression in a Four Bar Knee Joint on EM G Signals for an AK Amputee initia l setting of the assistant spring , is presented in Table 1. The min imu m EM G activity at the right limb for (0.4, 0.7, 1.1, 1.4, 1.8, and 2.0 km/h ) walking speeds happened to be at initial assistant spring settings of (+2, 0, 0, +2, +2, and +6 mm) respectively. While the minimu m EM G act ivity at the left limb and for the same above walking speeds happened to be at initial assistant spring settings of (+5, +6, +5, +6, +6, and +4 mm) respectively. It is clear that the EM G activ ity of the amputated limb is, in general, less than that of the intact limb and it is less fluctuating as the initial setting of the assistant spring varies. This indicates that, in order to achieve a stable movement by the amputee, mo re effort is exerted by the intact limb as compared to the amputated limb. Figure 4. Variation of applied force on the assistant spring at the standard initial setting with percentage of gait cycle Figure 5. Variation of the accumulated potential for different initial spring setting with the walking speed American Journal of Biomedical En gineer in g 2013, 3(5): 99-106 105 Figure 6. Variation of mean EMG activity of amputated (right) limb at the tested speeds with the variation of initial spring setting Figure 7. Variation of mean EMG activity of intact (left) limb at the tested speeds with the variation of initial spring setting 106 Somer M . D. Nacy et al.: Effect of Initial Spring Compression in a Four Bar Knee Joint on EM G Signals for an AK Amputee Table 1. Mean and standard deviation of EMG data (initial setting of assistant spring = +2 mm) Walking speed(km/hr) 0.4 0.7 1.1 1.4 1.8 2.0 Amputated (right) limb Mean (mv) 26.667 31.111 36.121 36.771 41.112 46.221 Standard deviation (mv) 2.027 2.488 3.173 3.126 3.453 3.998 Intact (left) limb Mean (mv) 34.111 36.889 41.122 43.222 46.666 50.333 Standard deviation (mv) 2.396 2.804 3.269 3.415 3.687 4.027 6. Conclusions pp: 50-57, 1-April-2002.  Liang Song, Xitai Wang, Siyuan Gong, Zengguang Shi, Fro m the results of this research, one can conclude that the Lingling Chen, "Design of Active Artificial knee Joint", effort exerted by the AK amputee person, at different Asian–Pacific Conference on M edical and Biological walking speeds, is great ly affected by the initial setting of the Engineering ( APCM BE ) ,vol:19, PP:155-157,2008. assistant spring in the four bar knee mechanism. Hence, as a  A. A. Torki, M . F. Taher, Abdalla Sayed Ahmed ,"Design future application, one can design and imp lement an And Implementation of A Swing Phase Control System for A electronic guided actuator to control the in itial setting of the Prosthetic Knee", Biomedical engineering conference , assistant spring for different walking speeds, thus producing minimu m effort exerted by the amputee person. CIBEC, IEEE, vol:978,No:1,pp:1-4, – Cairo ,18-20 Dec-2008. ACKNOWLEDGMENTS The authors wish to thank the staff of Bio mechanics Laboratory, Faculty of physical Education at Babel  Alex Sadra Olivera de Cerqueira, Edward Yuji Yamaguti, Luis M ochizki, Alberto Carlos Amadio and Julio Cerca Serrao, "Ground Reaction Force and Electromyographic Activity of Transfemoral Amputee Gait: a Case Study", Brazilian Journal of Kinanthropometry and Human Performance, vol: 15, No: 1, pp: 16-26, 2013. University for their help with the experiments and valuable  Donald R. Myers, and Gordon D. M oskowitz , "Myoelectric advises. They also thank Assist. Prof. Dr. Ali j. Abd, Faculty Pattern Recognition for Use in the Volitional Control of of physical Education at Babel University for h is help in installing the software of the used programs and help ing us in Above-Knee Prostheses", IEEE Transactions on Systems ,M an, and Cybernetics, vol:11,No.4,April-1981. ma king the experimental tests. They finally thank the staff of  Croline O Keefe Honours Project NCPO, "Muscle Activation biomechanics laboratory, faculty of physical education at the Patterns in the Transtibial Amputee", M onash Rehabilitation University of Baghdad, Jadiriya and specially Prof. Dr. Technology Research Unit, Rehab Tech, Australia, 1998. Sareih A. AL-Fad ly.  Dewen Jin, Ruihong Zhang , Jishuan Zhang , Rencheng Wang and William A. Gruver, "An Intelligent Above Knee Prostheses with EM G Based Terrain Identification", EEE International Conference,vol:3,pp:1859-1864,2000. REFERENCES  A. Bar, G. Ishai, P. M eretsky and Y.Korent ,"Adaptive M icro Computer Control of An Artificial Knee in Level Walking", Journal of Biomedical Engineering,vol:5, pp: 145-150, Ap ril-1983.  J. S. RIETM AN, K. POSTEM A and J. H. B. GEERTZEN, "Gait Analysis in Prosthetics: Opinions, Ideas and Conclusions" Prosthetics and Orthotics International, vol: 26,  E.C. Wentink , S.I. Beijen, H.J. Hermens, J.S. Rietman and P.H. Veltink, "Intention detection of gait initiation using EM G and kinematic data", Elsevier Journal of Biomechanics Gait and Posture, 2012.  Lingling Chen, Peng Yang, Linan Zu, and Yanli Geng, "The influence of walking speed on muscle activity of thigh and application in prostheses control",36th Annual Conference on IEEE Industrial Electronic Society,IECON,pp:2714-2718,710-nov.-2012.
... pages left unread,continue reading
Free reading is over, click to pay to read the rest ... pages
0 dollars，0 people have bought.
Reading is over. You can download the document and read it offline
0people have downloaded it