Bayesian probabilistic modeling in robosoccer environment for robot path planning

Abstract

The main goal of a route planning approach is to find a trajectory that safely transports the robot from one site to the next. Furthermore, it should provide an energy-efficient path so the computer can calculate it rapidly. This study develops a path-planning system for robots to approach the ball without collision. The Bayesian optimization algorithm (BOA) is used to identify the shortest path between the robot and the ball. BOA employs a probabilistic model to seek the optimum of an uncertain objective function efficiently. The performance of the BOA-based path planning system is compared to other optimization algorithms such as genetic algorithm, ant colony optimization, and firefly algorithm. BOA’s acquisition functions such as expected improvement, probability of improvement (PI), and upper confidence bound, are investigated. The exact locations of the robots and the ball are fed into optimization problems to discover the optimum path. The results reveal that the BOA system outperforms other systems in terms of computational time for planning the optimum path in dynamic situations and BOA-PI is the fastest algorithm.