Heuristic Parameter Tuning Procedures for a Virtual Potential Based AUV Trajectory Planner

This paper deals with the study and analysis of heuristics and tradeoffs incipient in the choice of method-independent parameters in an algebraic trajectory planner based on virtual potentials. The real-time trajectory planning framework for autonomous underwater vehicles (AUVs) is described in detail in Barisic et al. (2007), Barisic et al. (2008) and based on the referenced, as well as previous work of the authors. Of special interest among the parameters are those governing the strength of repulsion of obstacles (Barisic et al. (2007), and that of the strength of the rotary field (Barisic et al. (2008), Healey (2006)) that is introduced in order to eliminate obstacle-attached local minima present in classical virtual potential based methods (Healey (2006)).