Research On Collision-Free Motion Planning Of Agricultural Robot Manipulator

Agricultural robot, which was used widely in developed countries could help to mitigate the shortage of agricultural labor force, improve agricultural conditions with high productivity. Still, there were some problems such as slow searching speed, poor obstacles avoidance ability which can not be solved by traditional algorithms in practice.Aiming at realizing fast collision-free moving of agricultural robot, this thesis emphasize on collision-free motion planning of a 6-DOF agricultural robot manipulator.Firstly. the kinematics and dynamics models of 6-DOF robot manipulator MT-UARM were deduced, in which D-H method was adopted for kinematics equation and second Lagrange equation was employed for dynamics equation. Direct kinematics, inverse kinematics, bizarrerie positions and Jacobian matrix were analyzed. And redundancy of the inverse kinematics equation was resolved by genetic algorithms. According to the dynamics model, the moments of robot manipulator's joints were given. The work discussed above is the theoretical foundation of motion planning for robot manipulator.Secondly, general problems in robot manipulator motion planning were analyzed including trajectory planning and path planning. Several general methods of trajectory planning and path planning were introduced, compared and estimated in the application. The main work are the principles of time-optimal trajectory planning of robot manipulator, genetic optimization theory, and the algorithm of B-splines, based on which fourth-order quasi-uniform B-splines was adopted to construct trajectory. In result, a scheme of improved adaptive genetic algorithm (IAGA) was proposed based on population rally-and-dispersion of time-optimal trajectory planning. The physical constraints include joint velocity, acceleration, jerk and torques in the optimization. Simulation of B-splines path optimization in joint space for PUMA560 robot pre-3 joints was presented for illustration.Thirdly, the collision-free path planning problems for agricultural robot manipulator in obstacles environment were analyzed. Because of the complexity of working environment, collisions between arm and obstacles may occur. Therefore, it is necessary to program a collision-free path for the end-effector of agricultural robot manipulator, based on collision-free of all joints. The traditional artificial potential field method was modified, and a new method to eliminate the local rest point was proposed, which make the manipulator move at the direction of the least potential field. Combined with new constraint potential field function, realization process was given. And simulation of collision-free path planning was studied.Finally, simulation for MT-UARM 6-DOF robot manipulator in obstacles environment was studied, and its time-optimal trajectory planning based on B-splines was given. Simulation results verify the feasibility and accuracy of these methods.