A hybrid path planning system combining the A*-method and RBF-networks
In this paper we propose a novel, hybrid path planning system based on an extended A*-method in combination with special RBF-networks. The output of the A*-method, a set of classified cells, is used to train two variants of RBF-networks. Global RBF-networks (GRBF-networks) represent a wide area around the optimal path and generate smooth paths. Local RBF-networks (LRBF-networks) represent a small area around the optimal path and guarantee an obstacle-free “tube” surrounding this path. GRBF- and LRBF-networks are tested in different 3D- and 6D-scenarios.
Unable to display preview. Download preview PDF.
- 1.Kaspar Althoefer and Guido Bugmann. Planning and learning goal-directed sequences of robot-arm movements. In Proceedings of the International Conference on Artificial Neural Networks (ICANN'95) Paris, volume 1, pages 449–454, 1995.Google Scholar
- 2.Thomas Frontzek. Entwicklung eines neuronalen Bahnplaners mit heuristischer Vorverarbeitung. Master thesis, University of Bonn, Dept. Computer Science, 1996.Google Scholar
- 3.David Gelperin. On the optimality of A*. Artificial Intelligence, 8:69–76, 1977.Google Scholar
- 4.P. E. Hart, N. J. Nilsson, and B. Raphael. A formal basis for the heuristic determination of minimum cost paths. In IEEE Transactions on Systems, Man, and Cybernetics, volume 2, pages 100–107, 1968.Google Scholar
- 5.Tomaso Poggio and Federico Girosi. A theory of networks for approximation and learning. A.I. Memo No. 1140, MIT, 1989.Google Scholar
- 6.Charles W. Warren. Fast path planning using modified A* method. In Proceedings of the IEEE International Conference on Robotics and Automation, pages 662–667, 1993.Google Scholar