Smooth Path Planning For Dual-Manipulator Cooperation

Since the first industrial robot was invented,it has been developed into intelligent automation equipment and widely used in processing and manufacturing,etc.With greater demands are being placed on the productivity,the research of application of multi-robot cooperation is showing its importance.In this thesis,the research object is a working agent composed of two manipulators sharing workspace that overlaps in some region.Thus,collision free path planning and path smoothing is need for safe and stable operation of robots.For path planning for dual manipulators is a high dimensional planning problem,to realize high-efficient path planning and get smooth paths that are high quality and have small curvature,this thesis conducts the following researches based on heuristic search algorithm:construction of flat search graph and graph search method;design of heuristic function;path smoothing based on the optimization theory of B-splines.1.To overcome the shortage of inefficiency of heuristic search for three dimensional path planning,the method of constructing 2D planar search graph in 3D space is proposed.Thus the three-dimensional path planning of industrial robots is lowered to two-dimensional planning problem.To avoid sharp corner that the angle of path nodes is acute angle,the approach of exploring graph using search angle and radius is proposed.By setting the range of the angle to decrease the amount of exploring nodes when planning the path.The experiment results show that the proposed approach of graph construction and exploring reduces the amount of exploring nodes and computation time.Meanwhile,there's no sharp corner appearing in the paths.2.For collision free path planning,collision detection is realized by using sphere swept volumes to approximate the links and calculating the distance among these volumes.When planning the path,the informative distance from current position to the goal position and the minimum distance between two manipulators is combined to formulate the heuristic function.Meanwhile,the proportion of these two parts is evaluated by weight determined by the state of the working agent to balance collision avoidance and moving the manipulators.Therefore,the purpose of the design of heuristic function is to move the robots to the goal positions and obtain high path clearance.The results show that in the paths generated by the heuristic function design in this thesis,the path clearance is about 0.190~0.200?.3.Optimal B-splines based method is utilized to smooth the paths to obtain?~?continuity and curvature-constraint-satisfying paths.De Boor algorithm is used to compute the value and derivative of B-splines and the second derivative that's for smoothing and the approximation error are used to formulate the optimization objective function to get the optimal control points.The linear boundaries are determined to limit the valid domain of control points and upper bound curvature constraint is applied to formulate non-linear inequality constraint,which make the paths have small curvature and keeps validity of the paths and high path clearance.The results show the curvature of smoothing paths is smaller than the curvature upper bound and it change smoothly that make the paths have?~?continuity.The experiment results demonstrate the proposed path planning and smooth algorithm are capable of generating high quality paths for dual-manipulator agent.The paths are on planes,have?~?continuity,satisfy the upper bound curvature constraint and keep high path clearance.Therefore,it can ensure the safety of robots,decreases oscillation of joints velocity and acceleration,increase fluency and stability when the manipulators are moving.