Zone Model Predictive Control for Upper-Limb Exoskeleton Based on Improved Harmony Search Algorithm

Scientific and technological advancements have boosted automation and intelligence in industrial production, certain industries such as construction, manufacturing and logistics still rely on human labor. The upper-limb exoskeleton which designed to augment upper-limb strength, plays a significant role in assisting workers with tasks such as handling and sorting. It effectively reduces muscle fatigue and energy expenditure for the wearer. Aiming at the challenge of controlling accuracy and real-time performance of upper-limb exoskeletons, a zone model predictive control (ZMPC) method based on fast adaptive harmony search (FAHS) algorithm was proposed. Firstly, the optimization objective function of ZMPC, initially in quadratic programming (QP) form, was converted to a linear complementary problem (LCP). Secondly, solved this LCP by FAHS, which is an improvement of HS. We improved the initialization of the HM step and the generation of new harmony step. Finally, experimental results verified the superiority and effectiveness of the proposed control method.