Robot Joint Module Optimization Robust Algorithm Design With Simulation and Experimental Validation

ABSTRACT

Considering the presence of modeling errors, load variations, and time‐varying parameters in collaborative robot joint modules, we propose a novel robust approximate constraint‐following control algorithm based on the Udwadia–Kalaba (U–K) framework. The core innovation of this work lies in the analytical decomposition of the control law into three distinct components: A nominal term that governs the baseline dynamics, a correction term to address initial condition incompatibility, and a robust term designed to compensate for uncertainties. This structure enables constraint compliance and robustness without relying on online parameter estimation. To validate the proposed controller, we carried out both numerical simulations and real‐time experiments on a joint module test platform equipped with two representative friction models, using the fast control prototype system CSPACE. Simulation results indicate that the step response stabilization time is approximately 1 s, while the sinusoidal trajectory tracking error remains below 0.01°. Theoretical analysis and experimental results consistently demonstrate that the proposed method significantly improves the dynamic performance and robustness of the joint module system.