A Rapid Recursive Experimental Approach to Structure/Control Re-Design

Abstract

Abstract This paper introduces an experimental recursive method for simultaneously changing both the mechanical structure and control design of mechatronic systems in order to improve the system’s overall performance. The method improves a system’s closed-loop control specifications through recursive concurrent structure reinforcement and control gain optimization. By using a process of structural reinforcement, a single prototype structure can be used repeatedly until the system performance goals are achieved. To determine the optimal incremental structure changes, a recursive algorithm based on a gradient descent method and a parameter estimation theory is employed. After the incremental structure reinforcements are applied, the control parameters are optimized with respect to multiple control specifications. Next, the resulting system incorporating the structure and control changes is tested and compared with the desired level of performance. The entire process consisting of experimental evaluation, data analysis, and structure reinforcement is repeated until the system performance achieves the desired level. Simulation experiments are successful in changing both the structural and control parameters of a simplified positioning system and show improvement in the system’s overall settling time.