Advanced techniques, such as those referenced above, have rarely transferred to clinical practice, especially in the case of stroke rehabilitation, where the strategies adopted are either open-loop, or the stimulation is triggered using limb position or Electromyographic (EMG) signals to provide a measure of the participant’s intended movement. Closed-loop control has been achieved using EMG but this has not been incorporated in model-based controllers since EMG does not directly relate to the force or torque generated by the muscle. In the few cases where model-based control approaches have been used clinically, they have enabled a far higher level of tracking accuracy. The reasons for this are discussed in Freeman et al. (2012) with supporting references.

A principal reason for the lack of model-based methods finding application in a program of patient trials is the difficulty in obtaining reliable biomechanical models . In the clinical setting there is minimal set-up time, reduced control over environmental constraints and little possibility of repeating any one test in the program of treatment undertaken; control laws are required to perform to a minimum standard on a wide number of subjects and conditions. Moreover, the underlying musculo-skeletal system is highly sensitive to physiological conditions, including skin impedance, temperature, moisture and electrode placement, in addition to time-varying effects such as spasticity and fatigue (Baker et al. 1993). These problems are often exacerbated in the case of stroke because hemiplegic subjects exhibit both voluntary and involuntary responses to applied stimulation. The small number of model-based approaches that have been used in stroke rehabilitation therefore provide limited scope to adapt the applied stimulation to changes in the underlying system due to fatigue or spasticity, leading to reduction in performance and an inability to fully exploit the therapeutic potential.

This monograph describes the application of ILC in stroke rehabilitation, including clinical trials that constitute the first major stage towards eventual transfer into practice. In contrast to the other approaches employed to control FES, ILC exploits the repeating nature of the patient’s tasks in order to improve performance by learning from past experience. By updating the control input using data collected over previous attempts at the task, ILC is able to respond to physiological changes in the system, such as spasticity and the presence of the patient’s voluntary effort, which would otherwise erode performance. Use of ILC can also closely regulate the amount of stimulation supplied, ensuring that minimum assistance is provided, thereby promoting the patient’s maximum voluntary contribution to the task completion. As the treatment progresses this control action encourages patients to exert increasing voluntary effort with each trial, leading to a corresponding decrease in the level of FES applied.

The first research in this area concentrated on a planar problem that replicates the every day task of reaching out across a table top to, for example, a cup, where the aim was to establish the basic feasibility of using ILC in this setting. Figure 3.1 shows a stroke patient using the system designed for this purpose. More complicated tasks, such as reaching out and then lifting the arm are described in later chapters.

The patient in this figure is seated with her impaired arm supported by the robot and elliptical trajectories are projected onto a target above the hand. Also FES is applied to her triceps, using the surface electrodes, in order to assist tracking of a point that moves along the reference trajectory. At the end of the task, the arm is returned to the starting position in preparation for the next trial. During the reset time, plus a rest time to prevent muscle fatigue and allow transients to decay, an ILC law is used to calculate the stimulation to be applied on the next trial. The stimulation applied to the triceps muscle produces a torque about the elbow and the control problem is equivalent to controlling the angle in this figure. The shoulder strapping is to prevent forward movement by the patient’s trunk during the trials, which would conflict with the desired objective of reaching out with the arm.

Figure 3.2 shows a plan view of the patient’s movement in the planar case where the analogy with the pick and place operation for an industrial robot discussed in the previous chapter is clear. During the arm resetting time at the end of trial k, the ILC law uses a biomechanical model of the arm and muscle system, along with the previous tracking error, to produce the control signal, i.e., the FES, for application on the next trial.

The assessment of the results from clinical trials in this area must be based on the measures used by healthcare professionals. These are described below and the next chapter describes the ILC law design for this task and gives the results from a clinical trial.