A directional diffusion algorithm on cellular automata for robot path-planning

  • Authors:

  • Fabio M. Marchese

  • Affiliations:

  • Dipartimento di Informatica, Sistemistica e Communicazione, Università degli Studi di Milano-Bicocca, via Bicocca degli Arcimboldi 8, I-20126 Milan, Italy

  • Venue:
  • Future Generation Computer Systems - Cellular automata CA 2000 and ACRI 2000
  • Year:
  • 2002

Quantified Score

Hi-index 0.00

Visualization

Abstract

Cellular automata (CA) model is a powerful instrument used in many applications. In this paper we present a reactive path-planning algorithm for a non-holonomic mobile robot on multilayered cellular automata. The robot considered has a preferential motion direction and has to move using smoothed trajectories, without stopping and turning in place, and with a minimum steering radius. We have implemented a new algorithm based on a directional (anisotropic) propagation of repulsive and attracting potential values in a multilayered cellular automata model. The algorithm finds all the optimal collision-free trajectories following the minimum valley of a potential hypersurface embedded in a 4D space, built respecting the imposed constraints. Our approach turns out to be distributed and incremental: whenever changing the initial or the final pose, or the obstacles distribution, the automata start evolving towards a new global steady state, looking for a new set of solutions. Because it reacts to obstacles distribution changes, it can be also used in unknown or dynamical environments in combination with a world modeler. The path-planning algorithm is applicable on a wide class of vehicles kinematics, selected changing a set of weights.