An orientation assignment heuristic to the dubins traveling salesman problem

Abstract

In this paper we deal with the DTSP, which is the optimization problem where a path that goes through a set of two-dimensional points must be calculated considering the use of robots modeled as Dubins vehicles. Assuming that the sequence of visits is initially obtained accordingly to the ETSP, we propose an heuristic to assign orientations for each point in order to achieve a path which is length minimized and respects the vehicle’s nonholonomic constraints. The heuristic takes into account the vehicle’s minimum turning radius and distance between neighbors points to proportionally adjust the orientation on each point, allowing the definition of shorter Dubins curves connecting them. The methodology was horoughly evaluated through numerous trials in different simulated scenarios, providing statistical examination of the final results.