Composite Models for Mobile Robot Offline Path Planning
Ellips Masehian, M. R. · 2007
This chapter introduces two new offline path planning models which are based on the Roadmap and Potential Fields classic motion planning approaches. It is shown that how some relatively old methods can combine and yield new models. The first path planning model is established based on two traditional methods: the Voronoi Diagrams and Potential Fields, and so is called V-P Hybrid model. The model integrates the advantages of Voronoi diagram's safest distance and Potential Fields' search simplicity properties. After constructing the Generalized Voronoi Graph roadmap for the workspace, it is reduced to the Pruned Generalized Voronoi Graph (PGVG) through a pruning procedure. The PGVG decreases the search time effectively. An attractive potential is then applied to the resulting roadmap, which yields a new version of Potential Fields method, since it implicitly models the obstacles by attractive potentials rather than repulsive ones. The search technique developed for finding the trajectory is a bidirectional steepest descent ? mildest ascent stage-by-stage method, which is complete, and performs much faster than the classical Potential Fields or Dijkstra's methods.