Dynamic Scene Mapping And Trajectory Planning Of Manipulator In Unknown Environment

Joint trajectory planning is the key technology of manipulator.It plays an important role in the fields of life service,working in a new environment,underwater exploration and so on.In such tasks,the robot system is required to have the ability to perceive the external environment and obstacle avoidance.The system is generally composed of environment mapping,eye-hand calibration,trajectory planning,and collision detection.The environmental mapping and collision detection depend on accuracy of algorithm.Trajectory planning includes path planning and path point interpolation.Ensuring the completeness of planning,this paper studies implementing the representation of workspace,improve the efficiency of re-planning,and optimize the running time of the system.The open source planning algorithms use C-Space sampling method usually.These algorithms get different planning results in different experiences,and not consider the replanning.With the improvement of the quality requirements,the time complexity of algorithm increases exponentially.This paper has researched the following aspects:A inverse kinematics solution algorithm for non-spherical wrist manipulator is proposed.In this method,position and rotation are calculated separately.Analytical and numerical optimization method are combined to solve various groups solutions in the flexible space of the manipulator.Combined with the depth camera realsense,a "eye-in-hand" system is built and calibrated.The depth map is used to calculate the point cloud,and the through filter and outlier removal algorithms are combined to preprocess the point cloud.In order to enable the planning algorithm to repeatedly access the environment map,the memory is applied in advance to represent the grid map.GPU is used to maintain the map and calculate the large matrix,dynamically map the work space,and realize the real-time reconstruction of the working scene of the manipulator.In this paper,the typical path planning methods are analyzed,and the advantages and disadvantages of each method are compared.Finally,the D*algorithm with good re-planning performance is used to realize the path planning with the shortest path at the end of the manipulator in Cartesian space.Through interpolation and smoothing of the path points,the executable joint trajectory on the time axis is obtained.