利用姿态传感器多点位控制机器人交互方法

Multi-Point Control Robot Interaction Method Using Attitude Sensor

由于已有机器人交互方法的人体姿态识别结果不准确,导致操作完成时间和误差上升。为了降低识别误差、缩短操作完成时间,研究了利用姿态传感器进行多点位控制机器人交互方法。为了准确识别人体部件,利用构建的检测器选取身体躯干候选点。同时兼顾人体部件的位置和方向,获取最优区域,完成人体姿态识别。在上述基础上,设计了机器人交互多点位操作系统,实现了通过姿态及语音远程控制机器人的目的。采用姿态传感器及时反馈机械臂的工作状态和周边的虚拟现实环境,最终实现机器人交互多点位操作控制。实验结果表明,所设计系统能够准确识别人体姿态信息,实现了降低操作完成时间和误差的目的。

Due to the inaccuracy of the human posture recognition results of the existing robot interaction methods,the operation completion time and error increase.In order to reduce the identification error and shorten the operation completion time,the multi-point control robot interaction method was studied.To accurately identify human components,body trunk candidates were selected using the constructed detector.At the same time,taking into account the position and direction of human body parts,the optimal region was obtained to complete the human posture recognition.On the basis of the above,the robot interaction multipoint operating system was designed to achieve the purpose of remotely controlling the robot through human posture and voice.The attitude sensor was used to feedback the working state of the manipulator and the surrounding virtual reality environment in time to realize the interactive multi-point operation control.The experimental results show that the designed system can accurately identify the human posture information and reduce the operation completion time and error.