Towards Self-Tuning Lower Limb Prothesis

Available powered knee prostheses in the market uses a finite state impedance control to operate the prosthetic knee. Impedance control parameters are manually calibrated with trial-and-error method in order to reduce gait deviation. Such an approach is tedious, time consuming, and impractical. The research is aimed to design an intelligent control system that is capable of calibrating the impedance parameters for a semi-active powered prosthetic knee dynamically and automatically. The research consists of three major stages. The first stage investigates the gait activities of daily living for capturing user intent. The gait cycle of such activities in this case are based on direct inertial measurement unit sensors (IMUs) which consist of accelerometers and gyroscopes, processed by an event detection algorithm. This serves as the inputs to the intelligent control algorithm. The second stage includes the design and development of an intelligent control system adapted with impedance control strategy. The main intelligent control algorithm is based on an adaptive neuro-fuzzy inference system (ANFIS) with genetic algorithm for learning in order to tune towards the desired impedance output to control the prosthesis. A second algorithm is designed to monitor and correct the gait symmetry while acting as a feedback to the main control algorithm. The final stage includes a complete integration of the designed and developed system. NI myRIO is used as the prosthesis controller because of its capability of parallel processing is suited for such an approach. The designed system will then undergo further experimental trials and validation on a fixed test rig and also an amputee. Preliminary simulation results showed that the system is capable of calibrating the impedance control parameters. The prosthesis is also aimed at dynamically reducing the user's gait deviation. These initial results highlight the possibility for a dynamic and accurate impedance control parameter calibration, which can improve the control of lower limb prosthesis