PATH PLANNING FOR MULTIPLE TYPES OF AUTONOMOUS VEHICLES

Abstract

Various oceanic properties can be forecasted by numerical simulations. The forecast, however, is approximate as the boundary and initial conditions that determines the results of the numerical simulations are uncertain. Certain events, such as oil spills, require an accurate forecast of the oceanic properties in order to deploy an effective containment plan. The accuracy of the forecast can be improved by additional sampling in regions that affect the forecast. Due to limited resources, the path of the vehicles that carries out the additional sampling has to be planned to give maximum improvement in the forecast. A path planning algorithm for Autonomous Underwater Vehicles (AUVs), formulated using Mixed Integer Linear Programming, has been successful in generating a path plan for Monterey Bay in Jan 2003. The current interest is to deploy Autonomous Surface Vehicles (ASVs) and AUVs in a water channel in the Singapore Straits. Modifications made to the path planning algorithm for only AUVs are described. The results for the path plan generated by the modified algorithm shows a good coverage of the water channel where field experiments will be carried out in Jan 2009.