Tabu Search for the Surveillance Task Optimization of a Robot Controlled by Two-Dimensional Stochastic Cellular Automata Ants Model

Abstract

Robotics has been investigated by many researchers over the last few years for providing tasks at a lower cost which are often inhospitable to humans. This research proposes a stochastic model to solve the surveillance task, using a robot controller based on a cellular automata ants model. This way, the robot uses an ants inverted pheromone with a Tabu search to chose the next cell movement. The Tabu search consists in storing temporarily the last positions where the robot was and do not allow the robot goes to a repeated cell. So, the ants inverted pheromone will be used just with cells that was not visited in a certain time. For the next movement, the probability of the robot goes to a cell with a small quantity of pheromone is higher than the probability of it goes to a cell with a lot of pheromone. Many simulations in various scenarios and environments were performed, showing that by adding the Tabu search it was possible to improve the vigilant robot performance in relation to the precursor model investigated.