Evolving a path planner for a multi-robot exploration system using grammatical evolution

Abstract

Area exploration and mapping with teams of robots is a challenging application. As the complexity of this application increases so does the challenge of designing effective coordinated control. One potential solution to this problem is to explore some relevant parts of the design space automatically. In this paper, we present an approach which uses Grammatical Evolution to design a control function for coordinated path planning of teams of mobile robots. Simulation results are promising with evolved control functions showing performance better than handwritten control in term of amount of explored area.

Evolving decision-making functions in an autonomous robotic exploration strategy using grammatical evolution

“…In earlier works [7], [8], we demonstrated that evolutionary search can be used to successfully evolve numeric constants in a scoring function and show, in simulation, that the evolved functions can outperform handwritten code. In this work we extend the function search space substantially to include the structure of the scoring function.…”

“…A number of approaches in the literature either conform or could be made to conform to this general schema [1], [2], [3], [4], [5], [7], [8], [9]. For example, in Yamauchi's navigational framework [2] Env can be considered to current pose and distance fields; NOpts is a set of unvisited frontier cells; and Sigs are the distances to the frontier cells.…”

“…Our exploration strategy framework is adapted from our previous work in [7], [8] with few modifications to cater single robot exploration task. Furthermore, the exploration task has two objectives: maximising exploration and minimising power consumption; and one constraint: collision avoidance.…”

“…F sh is a global field that assigns each free cell in O with the distance to the nearest fixed obstacle, dist q ⊂ F sh [7]. dist q of an occupied or unexplored cell is set to zero.…”