Stroke is the largest cause of disability in developed countries, where a relatively small percentage of patients with upper-limb impairment following stroke regain full function. In particular, many of these patients experience difficulty performing everyday reaching and grasping tasks. Functional electrical stimulation (FES) can assist stroke patients in moving their impaired limbs and has been shown to increase upper-limb function. In addition, the benefits of FES are greatest when combined with maximal voluntary effort from the patient to perform the movement. This poses the problem of how to provide the correct level of FES to assist the movement with the requirement that maximal voluntary effort is also encouraged. In control systems terms an algorithm that directly regulates the input is required as opposed to one that adapts the controller.

The underlying premise of the research reported in this monograph is the use of ILC to regulate the FES applied during rehabilitation where the patient makes repeated attempts to relearn a task by repetition. In particular, the patient is presented with a reference, such as a lighted path to follow in reaching out over a table top to a cup, and attempts to follow it guided by a robot and with FES applied to the relevant muscle. During an attempt, the error between the reference and the trajectory generated by the patient is measured and once the attempt is complete the arm is returned to the starting position and this information is used by the ILC law to compute the FES to be applied on the next attempt. Use of ILC in this application is a technology transfer from industrial robotics to next generation healthcare.

In this monograph, the results of three programs of research are reported, including clinical trial evaluation which is essential to enable the eventual take up of this work by healthcare professionals. The first program (Chap. 4) focused on initial proof of concept by considering movement in one plane and stimulated one muscle group (triceps) to control movement around the elbow joint. Patients tracked a moving trajectory with their hand whilst FES was applied to assist with the movement. Following each trial, ILC updated the FES signal for the subsequent trial. Results showed improvements in tracking accuracy during the sessions. This initial research did not allow the patient to attempt to lift the affected arm and also movement in the plane was tightly constrained by the support.