Hybrid adaptive control for series elastic actuator of humanoid robot

Purpose Generally, humanoid robots usually suffer significant impact force when walking or running in a non-predefined environment that could easily damage the actuators due to high stiffness. In recent years, utilization of passive compliant series elastic (SEA) for driving humanoid's joints has proved capability many aspects so far. However, despite being widely applied biped robot research field, stable control problem powered by SEAs, especially process, is still challenge. This paper proposes model reference adaptive (MRAC) combined with back-stepping algorithm deal parameter uncertainties lower limb driven SEA system. an extension our previous (Lanh et al. , 2021). Design/methodology/approach Firstly, dynamic obtained. Secondly, since there are unknown and uncertain parameters model, Model Reference Adaptive Controller employed guarantee robust performance limb. Finally, experiment carried out evaluate effectiveness proposed controller mechanism. Findings effective can be humanoid-SEA Besides, effect coefficients law analyzed further improve response's quality. Research limitations/implications Even though simulation shows good results system response, practical not been implemented fully quality controller. Originality/value The MRAC process utilized combine external but maintain stabilization. simplified lower-limb proven appropriate taken future.