Manipulator Workspace Boundary Extraction and Its Application in Workspace Analysis of the Human's Upper Extremity

Abstract

Identification of a workspace and its boundary is an important issue in the design and performance evaluation for manipulators, especially for the exoskeleton rehabilitation robot. This paper estimates roughly the volume covering the workspace of a manipulator and then meshes the volume into a lot of small cubes. A distance-weight and a three-dimensional (3-D) distance-weight matrix are defined. The distance-weight describes the point distribution around the corners of 3-D cubes and the 3-D distance-weight matrix characterizes the degree of point distribution in the workspace. The smoothing procedure and boundary curve extraction of grey-scale images available in mathematic morphology are employed to process the 3-D distance-weight matrix. A general algorithm of the workspace boundary extraction is proposed. The set of boundaries of all slices of the workspace forms the manipulator's workspace boundary. Two examples are given to verify the workspace boundary extracting algorithm. The workspace boundary of the human's upper extremity is extracted and analyzed by using the proposed algorithm.