Abstract:Soft wires of robot operating tools often interfere with end-effector tasks in practical automated production scenarios. In order to avoid soft wires of industrial robot manipulators executing end-effector tasks, a real-time 3-D visual detection-based soft wire avoidance (3D-VDWA) scheme is proposed, which considers the soft wire detection and location, the soft wire avoidance, and the motion planning at the same time. The proposed 3D-VDWA includes three modules: 1) perception module; 2) motion planning module; and 3) soft wire avoidance module. The perception module is based on hue–saturation–value (HSV) range and depth information to detect and locate soft wires. The motion planning module is based on the method of the pseudo-inverse matrix to plan the path of the manipulator to the target position and orientation. The soft wire avoidance module is based on the Jacobian transpose method to dynamically avoid soft wire obstacles in the process of movement. Experiments demonstrate the effectiveness and the feasibility of the proposed scheme to solve the motion planning problem with soft wire avoidance of industrial manipulators for end-effector tasks.

A Real-Time 3-D Visual Detection-Based Soft Wire Avoidance Scheme for Industrial Robot Manipulators