Developing an adaptive controller for a shape memory alloy walking assistive device

The Shape Memory Alloy is a lightweight, compact and biocompatible actuation mechanism which is considered here to replace the current actuation technologies in assistive locomotion devices. This paper is aimed toward the development of an adaptive robust controller to deal with control problems in the actuation of shape memory alloys (SMA). In this research the ankle joint is considered to be actuated by SMA but it can be extended to the other joints as well. To check the performance of the controller, the dynamics of the ankle joint during walking is studied and an SMA manipulator with a similar behavior is used for the experiment. Nonlinear behavior of SMA wires requires nonlinear control techniques for tracking the desired ankle angle. Since the device is subjected to several uncertainties and unmodeled parameters, it is also necessary for the implemented control technique to be robust and adaptive. To this end, the proposed control technique consists of two parts. The first part is an adaptive PID controller which is motivated from a sliding mode control. The control gains are adjusted based on an adaptation mechanism to minimize the sliding condition. The second part of the controller is a supervisory control that guarantees the stability of the system.