Human Movement

Human movements are produced by contractions of muscles that develop enough forces to move limbs against the external loads. In a voluntary movement, a command signal from the central nervous system (CNS) is transmitted through the spinal cord to the muscle to develop forces. The normal muscle function requires an intact connections among the CNS, the spinal cord, and the muscle. Damage of the brain or the spinal cord injury (SCI) interrupts the command signal to reach the muscle. The human with brain damage or SCI have lost of motor control resulting in lose of the functional movements in daily life such as standing, locomotion, reaching. Types of the paralysis are classified as Monoplegia, Diplegia, Hemiplegia, Paraplegia, and Quadriplegia [1]. Definition of each type of the paralysis is given below:

(a) Monoplegia: paralysis of only one limb that is caused by isolated damage of the CNS or the peripheral nervous system (PNS)

(b) Diplegia: paralysis of same body region on both sides of the body, i.e. both arms or both sides of the face)

(c) Hemiplegia: paralysis of one side of the body. This paralysis is caused by damage of the brain. mainly caused by the damage of the cerebral palsy

(d) Paraplegia: paralysis of both lower limbs and trunk caused by the damage of the spinal cord

(e) Quadriplegia: paralysis of four limbs and trunk that is caused by the damage of the spinal cord.

In a person with the paralysis, the lost command signal from the CNS can be substituted by the artificial electrical stimulation to the peripheral nervous system or the muscle. This electrical stimulation acts in the same way as the electrical impulse from the CNS, resulting in muscle contractions and causing movements or sensations. This method is called the functional electrical stimulation (FES), with the aim of providing muscular contraction and producing a functionality of useful movement [2]. Clinically, FES is used as an orthotic aid with therapeutic effect. FES generates an indirect control of muscle contraction and movement and contributes to neurologically normalization of the impaired motor system due to the programmed and repeated pattern of the motor response [3]. Clinical researches pertaining to the FES involve restoration of some types of the human movement in the everyday activity such as grasping [4], [5], standing and walking [2], [5].

Gait is one of the cyclic movements. Each gait cycle is divided into two phases, the stance phase and the swing phase. In a certain sub phase of the gait, the movements of the joints reach certain joint angles (e.g. maximum knee flexion angle of swing phase, maximum ankle dorsifelxion angle of swing phase, ankle joint angle at initial contact). Gait cycle is a single sequence of events between two consecutive initial contact of the same limb. A single cycle of the gait is detected between two successive events of contact of the foot with the ground of the same leg.

The human movements induced by FES require the appropriate control method that can restore the desired functional movements. However, controlling the FES-induced human movements is difficult and complex due to the non-linearity of nuero-muscular system response [6]-[8], variability of response of the stimulated muscle [9], significant time delay, and muscle fatigue [7], [9]-[13]. Because of high inter-subject variability, in practical FES, the appropriate stimulation data should be identified on an individual basis of subject. The identification task needs a complicated experimental approach [8]. In surface FES, the identification task has to be performed at every rehabilitation session.

FES is expected to improve gait ability of the paralyzed gait. There are a large number of paralyzed gait subjects in the world. Basic gait ability of paralyzed subjects can be improved by foot drop correction and controlling of the swing phase. The stance phase can be improved by stimulating the knee extensor to stabilize the knee joint and by using the assistive equipment such as crutches. The Swing phase is important for forward propulsion during gait.

The movements of lower limbs during gait are complex multi-joint movements and involve interaction among the segments of the limbs. Restoration of the paralyzed gait using FES needs a sophisticated control strategy. Many researches proposed and tested certain control methods for the FES-induced gait.

Current Method of FES Gait Control

Many control methods of FES for lower extremities have been proposed in many researches. The methods are basically classified in open-loop [2], [5], [11], [14]-[17] and closed-loop control [18]-[22]. In current clinical application of FES for gait restoration, a manually triggered stimulator with the open-loop control is used. The open-loop control is simple and easy to be implemented in clinical application.

The FES systems that use the open-loop control can result in a good gait when the muscles do not fatigue and there are no external disturbances. In the basic openloop control, the muscle model is not used to predict the stimulation pattern. The initial stimulation pattern is basically determined from the muscle activation pattern from the EMG signal [17], trial and error [1], [13]. A musculo-skeletal model was utilized in generation of an optimized stimulation pattern through a computer simulation [15]. An open-loop controller that included an inverse model of the muscle that related the stimulation intensity and the desired position was tested [16]. Utilizing the model to predict the stimulation intensity was found to be more effective than one that did not. However, inaccurate tracking was shown due to the variability of the muscle torque that depended on the joint position. The authors concluded that the closed-loop control is needed in order to produce the accurate and repeatable position regulation. McNeal et al. [11] pointed that muscle fatigue (the quadriceps) limited the knee joint angle in achieving the desired knee angle and that the closed-loop control would be needed to achieve satisfactory performance without relying on support of upper limb to compensate the muscle fatigue. If the closed-loop control is not being used, a simple gain controller can be used to compensate the muscle fatigue [11]. A template stimulation pattern adopted from the EMG signal was tested to synthesize the paraplegic gait [17]. However, manual correction was needed to refine the stimulation pattern after evaluation of quality of the controlled gait during experiment. Trial and error setting and manual correction of the stimulation pattern during rehabilitation or experiment will be a physical and mental burden on the patient.

Theoretically, problem of the open-loop control can be compensated by adaptive control or by adding closed-loop feedback to correct error. The closed-loop control uses information of the controlled joint movements as feedback signals to regulate the electrical stimulation pattern. The regulations of the electrical stimulations are aimed to compensate variations of the musculo-skeletal system’s responses or the disturbances. Availability of the artificial sensor to measure output of the controlled system allows implementation of the closed-loop control. The adaptive controller can be realized by adjusting the controller parameter, or by using the model of the controlled system, which parameter of the model is adjusted in real time. The traditional closed-loop control utilized a desired trajectory of the joint angle to be followed by the controlled joint.

Kubo et al. tested feasibility of a PID controller for the knee joint angle [18]. The controlled knee joint movemenet was controlled to follow a desired sine curve. Result of the controlled knee joint angle. Although some delayed-responses of the controlled knee joint angle occurred, the controlled knee joint angle followed the desired sine curve smoothly without any oscillations.

Hatwell et al. developed a model reference adaptive controller for the knee joint angle [19]. The controller was a closed-loop version of the controller developed by Hausdroff et al. [16]. The controller was tested in controlling the swing leg in seated position of healthy and paraplegic subjects. The pulse width of the electrical stimulations of the hamstrings and the quadriceps were regulated by the controller. In the first test, the knee angle was severely oscillating and the maximum knee extension angle could not be reached. The oscillation was caused by rapid switch of the stimulation pulse widths due to inappropriate parameters of the model and the controller. Although tracking performance was improved after adjusting parameter of the model, oscillation occured occasionally.

Artificial neural network (ANN) is a computation tool that is recently used to capture characteristics of a certain system. Abbas and Chizeck designed neural network as a pattern shaper of the closed-loop controller of the knee angle controller [20]. The design was directly addressed to major problems of FES, customization of control system parameter for a particular individual , attaining resistance to mechanical disturbance, and real time adaptation to take account for the change of musculo-skeletal characteristics. Computer simulation test of the designed controller showed that by using the learned stimulation pattern, the generated knee joint movements could follow the desired trajectory. The change of characteristics of the controlled system due to the muscle fatigue and diversity among the individual subject could be compensated. However, the test was in simple single-joint control. Ideally, the controller should be tested in multi-joint movements to asses its feasibility. Experimental evaluation has not been performed, except for tracking muscle torque in isometric contraction [23].

The ANN is also utilized in a hybrid controller in associate to a controller, i.e., a proportional-integral-derivative (PID) controller or a fuzzy controller. In this configuration the ANN is as a feedforward controller, while the PID or fuzzy controller have a role as a feedback controller, formed a neuro-PID or neuro-fuzzy controller. A neuro-PID for controlling electrical stimulation of the quadriceps muscle was developed and tested experimentally in the healthy and paraplegic subject [21]. The controlled system was the

knee joint movements in a seated position as an approximation of the swing gait. The neuro-controller improved tracking performance of the PID controller. Chen et al. developed an ANN controller, a neuro-PID and a neuro-fuzzy controller for controlling the stimulation intensity of the ankle dorsiflexor to induce the ankle joint angle following a programmed trajectory [22]. The neuro-fuzzy controller had best tracking than the other controllers. However, these controllers were tested only for single channel stimulation. Although some advantages of utilization of the ANN and its combination with the PID and the fuzzy controller has been shown in both the computational and the experimental tests, training of the ANN takes a long time. This disadvantage obstructs the use of the FES system with ANN in real application.

Although the closed-loop control has been developed in some configuration of the controllers, it has not been used yet in the clinical FES gait because of difficulties of the identification and the prediction of characteristics of musculo-skeletal system. The appropriate generic controller that can identify and compensate the change of the characteristics of the musculo-skeletal system has not been found. Using the closed-loop control, repeatability of the restored joint movements has not been seen to be really feasible in long time. Success of some neuro-controllers shown in the preceding paragraphs were limited in single-joint control. Additionally, the training of the ANN to capture the dynamics of the pathological gait completely requires several types of pathological gait pattern and takes a long time.

The Cycle-to-Cycle Control

The cycle-to-cycle control delivers the electrical stimulation in the form of the open-loop control in a cycle of gait. Correction of the stimulation burst duration is given to a cycle of gait based on the evaluation of the previous cycle of gait. Controlling the paralyzed gait to follow target joint angles continuously is difficult to result in accurate response as shown in the literatures described in the preceding section. On the other hand the cyle-to-cycle regulations of the stimulation burst durations to achieve certain target joint angles seem to be easy to generate a successfull gait. Considering other method of the closed-loop FES gait control, the cycle-to-cycle control method is an candidate.

First implementation of the cycle-to-cycle correction of the gait was reported by Gracanin et al.[24]. The burst duration of the electrical stimulation of the peroneal nerve was determined by the patient’s swing time and regulated based on its value of the previous cycle by utilizing foot switch. Implementation of the multi-channel system was reported by Trnkoczy et al. in controlling the hip, knee, and ankle angles [25]. An example of the controlled joint angle trajectories is shown in Figure 1.2. The left figure was the controlled joint angle with FES and the right figure was without FES. Using this control method, the patient can choose his own gait cadence. The repeatability of the restored gait is also promising. The knee extension was shown to be possible to reach the maximum angle. However, the maximum values of other joint angles were not different from the un-stimulated joint angles. Targets of maximum joint angles were not explicitly defined in the control system. The regulations of the stimulation burst durations were not based on the feedback information of the obtained joint angles of the previous cycle.

More reliable system in implementing the cycle-to-cycle control was developed by Veltink [26] and Franken et al. [27] using a PID controller. The authors defined a target maximum joint angle to implement the cycle-to-cycle control in a closed-loop control scheme. The controlled maximum joint angle was delivered as feedback signal. The PID controller regulated the burst duration of the stimulation pulses of the current cycle based on the error of the previous cycles. The experimental tests were performed in control of the maximum knee extension angle [26] and hip angle range [27]. Regulation of the stimulation burst duration on the basis of the cycle-to-cycle control algorithm was shown to be feasible in the both literatures. Repeatability of the controlled movements was also confirmed. However, deterioration of the PID controller was found in compensation of muscle fatigue. Additionally, finding an appropriate parameter value of the PID controller for the cycle-to-cycle control was also a problem. Furthermore, these studies were in single-joint control.

In order to prepare the practical control for clinical use, the cycle-to-cycle control method has to be studied more extensive. This control method should be realized in controlling multi-joint movements. Since the non-linearity of response of the musculoskeletal system is the basic problem in FES control, the cycle-to-cycle control should be implemented in a controller that can compensate the non-linearity of the musculo-skeletal system response.

References

1. CHC Medical Library, http://www.chclibrary.org/micromed/000579790.html.
2. A. Kralj and T. Bajd, ”Functional electrical stimulation: standing and walking after spinal cord injury,”CRC Press, Boca Raton, 1989.
3. L. Vodovnik, T. Bajd, A. Kralj, F. Gracanin, P. Strojnik, ”Functional electrical stimulation for control of locomotor systems,” CRC Crit. Rev. in Bioengineering, pp. 63-131, 1981.
4. N. Hoshimiya, A. Naito, M. Yajima, and Y. Handa, ”A multichannel FES system for the restoration of motor function in high spinal cord injury patients: A respirationcontrolled system for multi-joint upper extremity,” IEEE Trans. Biomed. Eng., vol. 36, pp. 754-760, 1989.
5. Y. Handa, ”Current topics in clinical functional electrical stimulation in Japan,” J.Electromyogr. Kinesiol., vol. 7, pp. 269-274, 1997.
6. P. E. Cargo, P. H. Peckam, and G. B. Thrope, ”Modulation of muscle force by recruitment during intramuscular stimulation,” IEEE Trans. Biomed. Eng., vol. BME-27, pp. 679-684, 1980.
7. M. Levy, J. Mizrahi, and Z. Susak, ”Recruitment, force and fatigue characteristics of quadriceps muscles of paraplegics isometrically activated by surface functional electrical stimulation,” J. Biomed. Eng., vol. 12, pp. 150-156, 1990.
8. M. Ferrarin and A. Pedotti, ”The relationship between electrical stimulus and joint torque: a dynamic model,” IEEE Trans. Rehab. Eng. vol. 8, pp. 342-352, 2000.
9. A. Trnkoczy, ”Variability of electrically evoked muscle contraction with special regard to closed-loop control orthosis,” Ann. Biomed. Eng., vol. 2, pp. 226-238, 1974.
10. B. J. Bigland-Ritchie, F. Furbrush, and J. J. Woods, ”Fatigue of intermittent submaximal valountary contractions: central and peripheral factors,” J. Appl. Physiol., vol. 61, pp. 421-429, 1986.
11. D. R. McNeal, R. J. Nakai, P. Meadows, and W. Tu, ”Open-loop control of the freely-swinging paralyzed leg,” IEEE Trans. Biomed. Eng., vol. 36, pp. 895-905, 1989.
12. H. M. Franken, P. H. Veltink, M. Fidder, and H. B. K. Boom, ”Fatigue of intermittenly stimulated paralyzed human quadriceps during imposed cyclical lower leg movements,” Intl. J. Electromyogr. Kinesiol., vol 3, pp. 3-12, 1993.
13. S. A. Binder-Macleod, S. C. K. Lee, D. W. Russ, and L. J. Kucharski, ”Effects of activation pattern on human skeletal muscle fatigue,” Muscle & Nerve, vol. 21, pp. 1145-1152, 1998.
14. H. J. Chizeck, R. Kobetic, E. B. Marsolais, J. J. Abbas, I. H. Donner, and E. Simon, ”Control of functional neuromuscular stimulation system for standing and locomotion in paraplegics,” Proc. IEEE, vol. 76, pp. 1155-1165, 1988.
15. G. T. Yamaguchi and F. E. Zajac, ”Restoring unassisted natural gait to paraplegia via functional neuromuscular stimulation: A computer simulation study,” IEEE Trans. Biomed. Eng., vol. 37, pp. 886-902, 1990.
16. J. M. Hausdroff and W. K. Durfee, ”Open-loop position control of the knee joint using electricaal stimulation of the quadriceps and hamstrings,” Med. & Biol. Eng. & Comput., Vol 29, pp. 269-280, 1991.
17. R. Kobetic and E. B. Marsolais, ”Synthesis of paraplegic gait with multi-channel functional neuromuscular stimulation,” IEEE Trans. Rehab. Eng., vol. 2, pp. 66-78, 1994.
18. K. Kubo, K. Fujita, N. Itakura, Y. Iguchi, and H. Minamitani, ”Simultaneous closedloop control of knee and ankle joints using functional electrical stimulation,” IEICE Trans. Inf. & Syst., vol. J71-D, pp. 2197-2204, 1988 (in Japanese).
19. M. S. Hatwell, B. J. Oderkerk, C. A. Sacher, and G. F. Inbar GF, ”The development of a model target adaptive controller to control the knee joint of paraplegic,” IEEE Trans Automatic Control, vol. 36, pp. 683-691, 1991.
20. J. J. Abbas and H. J. Chizeck, ”Neural network control of functional neuromuscular stimulation system: Computer simulation studies, ” IEEE Trans. Biomed. Eng., vol. 42, pp. 1117-1127, 1995.
21. G.W. Chang, J. Luh, G. Liao, J. Lai, C. Cheng, B. Kuo, and T. Kuo, ”A neurocontrol of system for the knee joint position control with quadriceps stimulation,” IEEE Trans. Rehab. Eng. vol. 5, pp. 2-10, 1997.
22. Y. Chen, S. Chen, W. Chen, C. Hsiao, T. Kuo, and J. Lai, ”Neural network and fuzzy control in FES-assisted locomotion for the hemiplegic,” J. Med. Eng. & Tech., vol. 28, pp. 32-38, 2004.
23. J. J. Abbas and J. Triolo, ”Experimental evaluation of an adaptive feedforward controller use in functional neuromuscular stimulation systems,” IEEE Trans. Rehab. Eng. vol. 5, pp. 12-22, 1997.
24. F. Gracanin, A. KRALJ, and S. Rebersek, ”Advanced version of the Lubljana Functional Electronic Peroneal Brace with walking rate controlled tetanisation,” Proc. the 3rd Intr. Symp. on External Control of Human Extremities, Dubrovnik, pp. 487-493, 1969.
25. A. Trnkoczy, U. Stanic, and M. Malezic, ”Present state and prospects in the design of multichannel FES stimulators for gait correction in paretic patients,” TIT J Life Sci., 8, pp.17-27, 1978.
26. P. H. Veltink, ”Control of FES-induced cyclical movements of the lower leg,” Med. & Biol. Eng. & Comput., vol. 29, pp. NS8-NS12, 1991.
27. H. M. Franken, P. H., Veltink, G. Bardman, R. A. Redmeye and H. B. K. Boom, ”Cycle-to-cycle control of swing phase of paraplegic gait induced by surface electrical stimulation,” Med. & Biol. Eng. & Comput., vol. 33, pp. 440-451, 1995.

©arifin.tech@dmin