Solving an agricultural robot routing problem with binary particle swarm optimization and a genetic algorithm

Abstract

Agricultural robotics has become increasingly popular among agricultural researchers as an alternative to the use of human workers in the future. However, the operational cost of agricultural mobile robots must be competitive with the cost of hiring human workers. In agricultural mobile robot navigation, it is difficult to determine an optimized sequential route with a minimal distance. This paper employs binary particle swarm optimization (PSO) and a genetic algorithm (GA) to find the shortest routing path for spraying operations in a greenhouse. The agricultural robotics routing problem has been expressed in terms of the traveling salesman problem, which is commonly used in operational research. To solve the routing problem, an objective of a total path length was measured based on the path computed using a probabilistic roadmap path planner. The results indicated the performance of the GA was better for solution quality and computational time, while binary PSO performed better with respect to convergence time.