Optimal path planning with obstacle avoidance for autonomous surveying
David Hodo, David M. Bevly, John Y. Hung, Scott Millhouse, Bob J. Selfridge · 2010
In order to perform surveying with an autonomous vehicle, a path must often be designed for geometrically complex boundaries, while also accounting for mapped obstacles. In this paper, several algorithms that solve different aspects of the problem are presented. Together, the algorithms generate a path with the following characteristics: (a) it completely covers a field given its respective corner points, (b) considers the desired initial orientation of the vehicle, (c) provides methods for known obstacles to be avoided by defining areas that are not to be mapped, and (d) can cover a composition of multiple, disjoint fields. The algorithm represents an integration of several theoretical concepts, including: Dubins's paths for basic planning, modified visibility graphs to define feasible paths around obstacles, Dijkstra's algorithm to search for the shortest path, and...