An Adaptive Finite-Time Force-Sensorless Tracking Control Scheme for Pneumatic Muscle Actuators by an Optimal Force Estimation

Abstract

The prime challenges of the force control generated by pneumatic muscle actuators (PMAs) have arisen from their highly nonlinear properties, which are perturbed by dynamic uncertainties and external disturbances. To achieve fast response, high accuracy, and robust force tracking performance without using a force sensor, this letter proposes an adaptive finite-time force-sensorless control scheme for a single-joint manipulator configured by a pair of antagonistic PMAs. First, a newly time-varying adaptive optimal force estimation scheme is proposed to obtain the environmental contact force using only position information, such that improvements of transient response and estimated accuracy are assisted by the cost function of the predefined estimation error. Next, the principal control method relied on core elements of a finite-time integral fast terminal sliding mode control procedure is developed. Additionally, a force-based time delay estimation technique is deployed to estimate the lumped uncertainties. Meanwhile, the new adaptive gain law assists in compensating for the remained error of the estimator. The stability and finite-time convergence features of the whole system are demonstrated through Lyapunov theory. Finally, a series of trials are conducted on an actual paradigm with an adjustable environment configuration. The reliability and superiority of our approach are confirmed by comparing the experimental results with other approaches.