Collision Avoidance Path Planning For Manipulator In Complex Environment

Manipulator is widely used in industrial production,geographical navigation,material handling,danger detection,resource acquisition and other fields nowadays.It brings great help to people's work and lives.However,the application of manipulator is serious restricted by the complexity and variability of the work environment and the manipulator.Therefore,the research on collision avoidance path planning for manipulator is of great significance.A collision avoidance path planning method mainly involve environmental modeling,collision detection,path planning method.The exploration and research on collision detection path planning for manipulator has made some achievements,but there are still many problems to be studied.In this thesis,a research is carried out on collision detection path planning for six degrees of freedom manipulator with complex environment in our laboratory.Firstly,establishing the manipulator and environment objects models.For the problem from the complex shape,large quantity and real-time changing of the environmental objects and the manipulator,a tree object descriptive approach based on convex polyhedron models is proposed.The approach is characterized by a framework with a tree structure,besides the capabilities of precisely approximating the outline of an arbitrary object and mastering the DOF relationship among the affiliated bodies.A forward kinematics model for the tree structure object is realized based on the method,and the bounding-box model for the tree structure object is established.Next,researching the collision detection on manipulator itself or between manipulator and environment objects.Traditional collision detection method for manipulator is analyzed,because of the bounding box is composed by convex polyhedron,then traditional collision method of convex polyhedron is analyzed,then proposed a collision detection approach of convex polyhedrons based on the projection of edges,the experimental results show its high detecting accuracy,fast response speed and high reliability;for the continuous moving characteristics,an effective method is sampling the continuous path,building the bounding box between adjacent points,then a collision detection approach of manipulator in continuous motion is realized by the above method.Then,proposing a collision avoidance path planning method.The kinematics model of manipulator is established,many common targets on collision avoidance path planning is analyzed.Then transform the multi-objective optimization problem into a single-objective optimization by using weighting factors,and established fitness function to be solved easy by genetic algorithms.Then the path encoding and decoding path is designed,some key points on joint space is applied to path encoding and decoding path is carry out by linear interpolate between key points,then a collision-avoidance path planning method on joint space based on genetic algorithm is proposed.Finally,the collision avoidance work control system is designed and carrying out the experimental verification.An collision avoidance operation integrated system is established,a collision avoidance simulation and physical experiment is conduct using the system for manipulator in complex environment,the effectiveness and practicability of the relevant algorithms and the system is verified by the experimental results.