| It has become a trend to use robots in complex working environments to achieve multi-robot and even human-robot collaboration.Thus that the manipulator can perform tasks in a complex and changeable environment while ensuring that it does not interact with dynamic and static obstacles in task space is required.The traditional industrial robot,however,can only accomplish a single repetitive pipeline task by teaching programming,when the work environment changes,it needs to be reprogramed,so the adaptability is poor.Redundant manipulator is a kind of manipulator whose degree of freedom is greater than the task space.Its redundant characteristics make it more flexible and can accomplish many tasks that traditional industrial robots can not handle.But that also make it more difficult to plan.Therefore,it is of practical significance to study the dynamic obstacle avoidance planning of redundant manipulators.Most path planning algorithms can not be realized due to high computational complexity.However,artificial potential field and rapidly-exploring random tree do not suffer this problem.Therefore,in this paper,a redundant manipulator and its dynamic obstacle avoidance planning are studied.An improved algorithm is proposed based on the traditional artificial potential field and the rapidly-exploring random tree algorithm according to the characteristics of the redundant manipulator.The main research work of this paper is as follows:1.The control of manipulator is implemented in joint space,while environmental informa-tion is often acquired in cartesian space,so cartesian space needs to be mapped into joint space by inverse kinematics.But currently there are no unified algorithms to solve the inverse kinemat-ics of redundant manipulator.Therefore,some algorithms which are commonly used in inverse kinematics of redundant manipulators are studied and analyzed in this paper,including numerical solution and analytical solution,and the advantages and disadvantages of these algorithms are also introduced.In the path planning phase,the inverse kinematics solution methods should be chosen properly according to the characteristics of different planning algorithms.2.Aiming at the problem that the redundant manipulator has more than six degrees of freedom and can not construct the artificial potential field in the configuration space to achieve the goal of dynamic obstacle avoidance,an improved artificial potential field method is proposed.The attrac-tive velocity is directly defined at the end of the manipulator,while the repulsive velocity is defined at the nearest point between manipulator and obstacles,and the two velocities are synthesized in the joint space of the manipulator by using the inverse kinematics of the redundant manipulator,so that the artificial potential field can be used for high degree-of-freedom manipulator,and avoid constructing the configuration space.A Pivot algorithm is introduced to solve the collision problem when the velocity of the arm may be lower than the obstacle,and a virtual obstacle is constructed by connecting the nearest point and the target when the manipulator is trapped by the local minima.The simulation experiment result shows the effectiveness of the proposed algorithm.3.Although the artificial potential field method can realize the planning of the dynamic ob-stacle avoidance of redundant manipulator,the arm is easy to shake when it moves around the obstacle,and the trajectory can not be optimized because it can not be obtained in advance.To solve this problem,an improved rapidly-exploring random tree algorithm is proposed,including offline planning and online planning.The offline planning first uses the analytical solution of in-verse kinematics to determine the optimal target state with the artificial potential field,then builds the search tree with the target state as the goal,and obtains a collision-free path from the initial state to the target state,solving the problem that the rapidly-exploring random tree algorithm can not be used for online planning because of the computational complexity by constructing the search tree.The online planning switches the root node of the search tree according to the current state of the manipulator,then expands and rewires the search tree according to the current environment,and obtains a path from the current state to the target state in real time to avoid the dynamic obsta-cle.When the original target state is obstructed by the dynamic obstacle,the target is switched to re-search for a new path-The simulation experiment results show the effectiveness of the algorithm. |
