In the paper we present a robot arm somatosensory interaction manner based on the human skeletal information. First, we use Kinect sensor to capture image depth of field data, and deal with the data by skeletal tracking technology to match various parts of the human body followed by further establishing the 3D coordinates of every joint of human body. Then, we utilise the 3D coordinates to calculate the rotation angle of human waist, shoulder, elbow and wrist joints. These four joints of human body are in one-to-one correspondence with the four joints of 4-DOF robot arm. Finally, we transform the calculation results to control commands and sent them to controller of the robot arm by serial ports to realise somatosensory interaction. Experimental results show that the method can effectively control the robot arm to rotate and grab objects.%提出一种基于人体骨骼信息的机械臂体感交互方式.首先利用Kinect传感器获取图像景深数据,并通过骨骼追踪技术处理景深数据,以匹配人体的各个部分,进而建立人体各个关节的3D坐标.然后利用3D坐标值计算出入体腰、肩、肘和腕4个关节的转动角度,人体的这4个关节与四自由度机械臂的4个关节一一对应.最后将计算结果转换成控制指令通过串口发送给机械臂的控制器,实现体感交互.实验表明,该方法能够有效地控制机械臂的转动以及抓取物件.
展开▼
