A new algorithm for construction of discretized configuration space obstacle and collision detection of manipulators

Abstract

A novel approach for configuration space obstacle(C-obstacle) construction and collision detection of manipulators by precompiling the global topology of a robot's free space is proposed in this paper. The approach consists of an off-line phase plus an on-line phase. In the off-line phase, a C-obstacle database (COD) for a given robot is setup in which the C-obstacle bitmaps are stored and indexed by the cells of the workspace; in the on-line phase, when the same robot is operated in a real environment, those maps whose indexes match the real obstacle cells in the environment are identified and then extracted from the COD database. The superposition of these maps forms the required on-line C-obstacle in the application scenario. Based on the COD database, collision can also be detected for any given configurations of the robot in a dynamic environment. The proposed approach is a general one which is applicable to manipulators with any kinematic structure and geometric shape. Since the time-consuming construction of the COD database is implemented in off-line phase, the on-line running time is only for the identification of components in the COD database, hence the proposed approach for C-obstacle construction and collision detection is applicable for robot manipulation in dynamic environments. Simulations involving different types of manipulators are studied and the results have verified the feasibility and efficiency of the proposed approach