An approach to on-line path planning for multiple robots

Presents an approach to on-line path planning for multiple robots in configuration space. On the basis of defining some points in workspace as fundamental obstacles, the author develops an efficient algorithm for mapping a complex Cartesian obstacle by selecting the obstacle´s critical points. Its computational time for mapping a two-dimensional obstacle is approximately 5 ms with a 33 Mhz 80486 CPU. To plan a collision-free path for a master robot, the author uses slice configuration space obstacles for building a free subspace. To plan a collision-free path for a slave robot, swept volumes of the master robot are taken as additional moving obstacles into consideration. Finally, the author reports graphical simulation results with respect to two PUMA robots sharing a common workspace