Adaptive ankle impedance control for bipedal robotic upright balance

Upright balance control is a fundamental skill of bipedal robots for various tasks that are usually performed by human beings. Conventional robotic often realized developing accurate dynamic models using series fixed torque-ankle states, but their success subject to physical and kinematic models. This can be particularly challenging when external disturbing forces present, this common in unstructured working environments, leading ineffective control. To address such limitation, paper presents an adaptive ankle impedance method with the support advances fuzzy inference systems, which desired torques generated real time adaptively meet requirement. In particular, initialised specific first representing general situation, then evolves whilst performing real-world environment acting on feedback from system. implemented initialising rule base typical allowing evolve environments. closed loop action mechanism timely effectively configures system requirements. The proposed was applied robot moving vehicle validation evaluation, loads ranging 0 1.65 kg disturbances terms acceleration 0.5 1.5 m / s 2 $$ \mathrm{m}/{\mathrm{s}}^2 , swing angles up 7.6 º {7.6}^{\mathrm{^0}} anti-disturbance timespans 8.5 s. These experimental results demonstrate power upright improving robustness, thus applicability, robots.