Constrained Navigation with Mandatory Waypoints in Uncertain Environment

François Lucas (1) , Christophe Guettier (1) , Patrick Siarry (2) , Anne-Marie Milcent (1) , Arnaud de La Fortelle (3, 4)

1 SAGEM Défense Sécurité [Massy]
2 LISSI - Laboratoire Images, Signaux et Systèmes Intelligents
3 CAOR - Centre de Robotique
4 IMARA - Informatique, Mathématiques et Automatique pour la Route Automatisée

This paper presents a hybrid solving method for vehicle path planning problems. As part of the vehicle system architecture (vetronic), planning is dynamic and has to be activated on-line, which requires response times to be compatible with mission execution. The proposed approach combines constraint solving techniques with an Ant Colony Optimization (ACO). The hybridization relies on a static probing technique which builds up a search strategy using a distance information between problem variables and a heuristic solution. Various forms of this approach are compared and evaluated on real world scenarios. Preliminary results exhibit response times close to vehicle control requirements, on realistic problem instances.

Mots clés: Path planning; ant colony optimization; probing; TSP; constraint programming; search

Domaines: Sciences de l'ingénieur [physics]; Automatique / Robotique

François Lucas, Christophe Guettier, Patrick Siarry, Anne-Marie Milcent, Arnaud de La Fortelle. Constrained Navigation with Mandatory Waypoints in Uncertain Environment. International Journal of Information Sciences and Computer Engineering, 2010, 1 (2), pp.75-85. ⟨hal-00831578⟩