A Safe Admittance Boundary Algorithm for Rehabilitation Robot Based on Space Classification Model

Existing studies on rehabilitation robots are generally devoted to robot-assisted active training, which is conducive facilitating muscle and nerve regeneration. However, human–robot interaction (HRI) requires imposing a limit the workspace within robot operates, so as ensure patient safety. A safe admittance boundary algorithm for proposed based space classification model works by constructing virtual HRI in control layer. First, point cloud isodensification performed workspaces human body operate. Next, nearest neighbor density determined cloud, built this basis. Finally, integrated with derive algorithm, can be used safety control. This then subjected verification experiments out−of−bounds using dynamic arm simulator (DAS). As indicated experimental results, when side length of voxel grid center filtering set 0.06 m, accuracy definition 98.40%. The average maximum response time behaviors 165.62 ms. impose reasonable constraints workspace, thereby improving