摘要
针对复杂环境下仿人机器人稳定控制的要求,设计了基于MEMS惯性传感器的仿人机器人足部姿态实时感知系统。同时,提出了基于惯性传感器和足部力传感器信息的系统累积误差校正方法,用于解决由于噪声、漂移和积分算法而带来的累积误差问题。该系统能够实时精确的输出仿人机器人的足部姿态信息,为仿人机器人的稳定控制提供关键依据。该校正方法能够有效的消除系统的累积误差,提高系统的测量精度。最后在仿人机器人BHR-2上对该系统的可行性和有效性进行了实验验证。
This paper proposes a real-time foot attitude estimation system based on MEMS inertial sensors for the need of stabilization control of humanoid robots in complex environments. A cumulative error calibration method based on the inertial sensors and the force sensor equipped in the foot is presented synchronously in order to eliminate the influence by noises, shifts and the integral algorithm. The system can provide real-time and exact foot attitude information as crucial parameters of stabilization control. The cumulative error calibration method can reduce the influence of cumulative error and improve the precision of the system. Finally, the feasibility and validity of the system is confirmed by experiments with the humanoid robot ＇BHR-2＇.
出处
《微计算机信息》
2009年第23期168-170,共3页
Control & Automation
基金
基金申请人:黄强
项目名称:基于人体运动的仿人机器人运动设计方法
基金颁发部门:国家自然科学基金委(60874048)
关键词
仿人机器人
足部姿态感知
累积误差校正
惯性传感器
力传感器
Humanoid robot
Foot attitude estimation
Cumulative error calibration
Inertial sensor
Force sensor
