Research On Path Planning Of 6-DOF Manipulator Based On Improved RRT-Connect Algorithm

Path planning for manipulator is one of the most important research topics in robotics,and it is of great value.Scientific and reasonable path planning is a prerequisite for the normal and efficient operation of manipulator.As the requirements for automation and intelligence of manipulator increase in the course of industrial development,the original path planning algorithms can no longer meet the requirements.In order to improve the work efficiency of manipulator and to ensure that they can complete their tasks in a smooth and safe manner,it is important to study new path planning algorithms.The six-degree-of-freedom manipulator is one of the most widely used robots in the manufacturing industry.It has the characteristics of flexible operation,reliable work,free programming and high versatility,etc.Therefore,this paper takes the UR10 robotic arm as the research object,carries out kinematic modelling and analysis of the robotic arm,improves the original path planning algorithm,and builds a 3D simulation platform based on MATLAB and ROS,and carries out algorithm simulation verification on the simulation platform with the established robotic arm model as the research object.Firstly,according to the structural parameters of the UR10 manipulator,the D-H method was used to model the UR10 manipulator,and the positive and inverse kinematic equations of the manipulator were derived using the principle of spatial coordinate transformation.In order to check the correctness of the positive kinematic model,the Robotic Toolbox simulation toolbox in MATLAB was used for verification.Next,several commonly used collision detection models are analysed for the collision detection problem of the robotic arm in the path planning process,using cylindrical and spherical enclosing boxes to simplify the robotic arm and the obstacle respectively.For the shortcomings of lack of directionality and slow convergence in the expansion process of RRT* and RRT-Connect algorithms,a new RRT-Connect algorithm with the introduction of an intelligent sampling function is given based on the RRT-Connect algorithm,in addition to which the random tree search method in the new RRT-Connect algorithm is also improved.Simulations of RRT*,RRT-Connect and the given improved RRT-Connect algorithm are carried out in MATLAB for obstacle avoidance path planning to verify the effectiveness of the new algorithm.To address the problem that the smoothness of the paths planned by the new RRT algorithm is still unsatisfactory,a method of post-processing the new algorithm using B-sample curves is proposed,which is verified by simulation to be effective in increasing the smoothness of the paths planned by the new RRT-Connect algorithm.Finally,a complete motion configuration function package is obtained by combining the Solid Works exported six-degree-of-freedom manipulator model with the Moveit function package in ROS,the current mainstream international open-source robot operating system.The RRT-Connect algorithm and the RRT* algorithm were used for path planning in two different obstacle scenarios,to derive information on the motion state of the joints of the manipulator and to analyse the results obtained from the simulation.