基于多传感器的机器人遥操作人机交互系统

为了增强机器人人机交互系统的自然性，提出了基于多种传感器的非接触式人机交互系统设计方案,系统通过检测操作者手部动作和手部位置姿态的变化实现机器人的遥操作。研制了肌电传感器，获取手臂上一对拮抗肌上的表面肌电信号，并以此来判断机器人操作者的部分手部动作；利用Kinect体感设备和惯性测量单元获取手臂三维位置和姿态角信息。通过网络将人手的动作及位置姿态发送至机器人控制系统，以完成对机器人的控制。系统综合多种传感器的优点，极大减小了传统接触式交互方式对操作者运动范围的限制，实现了自然交互，实验表明了其有效性。

Man-Machine Interaction System for Robot Teleoperation Based on Multiple Sensors

In order to enhance the naturalness of common robot man-machine interaction systems,a design scheme of contactless man-machine interaction system based on multiple sensors is proposed,which implements the teleoperation by detecting the gesture and movement of operator''s hand.A kind of sEMG(eurface electromyograbhy) sensor was developed,which acquired the sEMG signals of a pair of antagonistic muscles to estimate the partial motion of hand.The positions of every joints of arm in 3D space were acquired through the Kinect device,and the information about the rotation angles of forearm were acquired by using an IMU(inertial measurenent unit).All the data of motion,position and rotation were sent via the Internet to the robot control system to complete the control of the robot.By combining the advantages of multiple sensors,the designed system greatly reduces the range restriction on operators in traditional interactive ways.It achieves a natural interaction,and the experimental results validate its effectiveness.