正弦驱动与传感反馈结合的双足机器人仿生行走控制

A Bionic Walking Control for Biped Robot Using Sinusoidal Input and Sensory Feedback

提出了一种正弦驱动与传感反馈结合的双足机器人仿生行走控制方法.所有关节由正弦振荡器驱动,较之相互耦合的神经元振荡器更加简单;控制参数具有明晰的物理意义,便于对运动模式进行调节.传感反馈表征了机器人的运动状态,对于保证机器人的稳定行走起着至关重要的作用.将机器人碰地、碰膝等关键运动状态作为相位反馈,对控制力矩进行相位重置,协调各关节动作,进而实现控制器、机器人、环境的耦合.同时,从节省能量和仿生的角度,考虑了关节运动的被动特性,确定了各关节力矩的作用区间.仿真结果表明,该控制方法能实现机器人稳定行走,并具有良好的能效性和自稳定性.

A bionic walking control strategy for biped robot using sinusoidal input and sensory feedback is proposed. All joints are actuated by sinusoidal oscillator, which is simpler than coupled neural oscillators. Control parameters have clear physical meaning, and it is convenient to adjust walking pattern. Locomotion states of robot are characterized by sensory feedbacks, which play an important role in ensuring stable walking. Some key locomotion states, such as ground impact and knee impact, are used as phase feedback to reset joint torques and coordinate movement of different joints. The coupling among controller, robot, and environment is then realized. The actuating areas of torques are determined considering passive characteristic of joints from the perspectives of saving energy and imitating human beings. The simulation results show that stable walking with favorable energy efficiency and self-stability is obtained using this control strategy.