Research Of Collision-free Trajectory Planning Algorithms For Industrial Robot

Collision-free trajectory planning is considered to be one of the most important problemsin trajectory planning for robots. The primary goal of collision-free trajectory planning forrobots is to acquire a trajectory from the start point to the target point without colliding withthe obstacles in their configuration space. The research presented in this paper is about theproposed Velocity Vector Field method, which is based on the traditional Artificial PotentialField method, and its application in dealing with the collision-free trajectory planningproblem. The organization of this paper is as follows.To begin with, forward kinematics and inverse kinematics analysis of the SCARA robotdiscussed in this paper are presented. A collision-free trajectory planning system has beenbuilt by using Simulink and importing robot’s SolidWorks models into SimMechanics. Bycombining the robot’s3D MAX geometric model and OpenGL, a motion simulation platformfor the SCARA robot has been established.Second, based on the traditional Artificial Potential Field method, which might causelocal minimum problem, a modified Velocity Vector Field has been proposed forcollision-free trajectory planning. The formulation and implementation of the Velocity VectorField Method are described. The method could be applied in avoiding moving obstacles andreal-time collision avoidance. Besides, the Velocity Vector Field was modified to generatemore stable velocity for the robot.Third, fuzzy logic method has been used to optimize the proposed method, which couldmake the Velocity Vector Field method more adaptive. Sliding mode variable structuremethod has been used to optimize the proposed method, so the distance between the robot andthe obstacle could be controlled better. Simulation and experiments results prove that the twoproposed methods are effective. The generated trajectories are then optimized so the velocityof the robot would become more stable.Finally, the trajectories generated by the MATLAB/SimMechanics trajectory planningsystem are applied on the OpenGL motion simulation platform, the effectiveness of theproposed methods has been proved by the simulation platform. Moreover, experiments havebeen performed on a SCARA robot and confirm the performance of the presented methods.Both the simulation and the experiment results verify the applicability of the researchachievement of this paper.