Path planning for information collection tasks using bond-energy algorithm

Autonomous vehicles are often employed to explore an area and collect information from objects of interest. In this paper, we consider path planning for an autonomous vehicle in a generalized information collection infrastructure. A novel information collection model that accounts for the vehicle visibility range is proposed which enables the modeling of non-isotropic targets, i.e., targets with unequal information emission rate with respect to the relative position of the vehicle. Finding a feasible solution for such a path planning problem can be challenging since the problem can be easily over-constrained. Moreover, existing methods, such as the many heuristics for Traveling-Salesman Problem (TSP), may not provide a feasible solution since they do not necessarily account for information-collection objective. The bond energy algorithm, an optimization heuristic that is robust to initial conditions in information collection tasks, is provided. Using these models and techniques, a path planning algorithm is proposed for an arbitrarily defined information collection mission. Several sample scenarios are demonstrated, one with objects with isotropic information emission, and the other with non-isotropic emission.