Minimally Model-Based Trajectory Tracking and Variable Impedance Control of Flexible-Joint Robots

Abstract

This article presents a novel control scheme that primarily provides a solution to the problem of simultaneously performing stable trajectory tracking and variable impedance control (VIC) of flexible-joint robots. The existing asymptotically stabilizing tracking controllers for compliant robots rely upon several model-based terms and their derivatives. Contrarily, the proposed minimally model-based trajectory tracking control scheme requires knowledge of only the gravitational torque vector model. Although interaction and tracking performance enhancements are achievable through VIC, its operation introduces energy injections, thereby impacting closed-loop stability/passivity. Existing VIC passivity preservation techniques deal exclusively with rigid-joint robots. To this end, the proposed method introduces the minimally model-based trajectory tracking variable impedance control scheme, which ensures passivity/stability when applying VIC to flexible-joint robots. Numerical and experimental results corroborate the controllers' efficacy in terms of realizing VIC interaction and tracking tasks, thus enabling flexible-joint robots to stably replicate biologically inspired behaviors.