Decentralized Adaptive Pseudo-Inverse Control for Output Constrained Systems on Giant Magnetostrictive Actuators

Rigid smart material actuators, such as giant magnetostrictive actuators (GMAs), are characterized by high energy density, fast response speed and large mechanical output. It is widely used in the fields of industrial robotics, precision positioning and microelectronic systems. This paper proposes a decentralized adaptive pseudo inverse control scheme based on neural networks for a class of nonlinear time-delay systems with output constraints, which reduces the impact of hysteresis on the system. The main research content includes 1) the pseudo-inverse control programme is proposed to overcome the hysteresis phenomenon in the actuator 2) with the help of RBFNN, finite coverage primitives and potential Lyapunov function, the output-constrained control problem of interconnected time-lagged nonlinear systems is solved by designing an approximator for the time-lagged nonlinear function; and 3) an experimental platform for precision control based on a three-axis GMA is designed, and the scheme is verified for the control performance.