Feedforward Compensation And Cascaded Control Scheme For Trajectory Tracking Of Pneumatic Muscle Actuated System

Over the past decade, pneumatic muscle actuators (PMA) has been steadily receiving much attention not only in the areas of industrial applications as well in promising research areas such as robotics and biomedical engineering. The popularity can be much associated with the attractive advantages PMA has to offer such as inherent compliant safety, high power to weight ratio and compact form factor. Despite the attractive advantages it has to offer, PMA exhibits significant nonlinear characteristics such as hysteretic behavior and creep phenomenon. Subsequently, these dynamic and time varying behaviors often makes modelling and real time motion control a challenging effort. Although many control methods have been developed, these controller design procedures frequently require exact model of mechanism and deep understanding in modern control theory which leads to their impracticability. Henceforth, in this research, a practical control strategy namely the Feedforward Compensation with Cascaded Control (FFC) scheme is proposed for the trajectory control of the PMA mechanism. The practical control scheme employed heavily considers on simple structure and straightforward design framework. Hence, the proposed FFC controller includes control elements that are derived from the measured open loop responses. The tracking performance is examined and compared to a Proportional Integral Derivative (PID) controller through experimental works. Experimental results show that the proposed controller can produce zero steady state error in step positioning. Similarly, the feedforward compensation with cascaded control scheme performs better in tracking when compared to PID controller with a higher tracking accuracy with an average improvement of 45 % and 64 % for maximal tracking error and root mean square error respectively. Likewise, when evaluated for robustness towards load variations, the proposed control strategy provides an ameliorated performance over the PID controller with an error improvement of 58 % in terms of maximal tracking error and 44 % in terms of root mean square error