Smooth, rigid, and dexterous robotic machining path planning based on on-site scanned point clouds

Abstract

Industrial robots with integrated sensing units possess the advantages of parallelism, dexterity, and intelligent operation, making them a cost-effective alternative to large and expensive machine tools for processing large and complex parts such as aircraft skins, ship hulls, and wind turbine blades. These complex parts are susceptible to deformation during transportation, clamping, and assembly. Therefore, employing robots to perform on-site scanning of clamped parts to acquire point clouds and subsequently conducting adaptive machining path planning based on these point clouds is a viable approach. In this paper, a novel method for planning a smooth, rigid, and dexterous robotic machining path based on on-site scanned point clouds is presented, which is suitable for the semi-finishing or finishing stages. First, a dual quaternion non-uniform rational B-spline (NURBS) curve fitting method is proposed to generate a smooth tool path from the point clouds. Then, to address the functional redundancy of the robot, a method is proposed to select the postures of all machining path points, optimizing the smoothness of the robot's joint space trajectory, rigidity, and dexterity under the constraints of joint limitations and avoiding collisions. A directed graph with robot configurations as its nodes is constructed, and an improved A* algorithm is proposed to find the shortest path in the directed graph. We conducted simulations and actual robotic cutting experiments, which demonstrated that executable robotic machining paths can be obtained, achieving a superior balance of joint space trajectory smoothness, rigidity, and dexterity.