In this article are compared random waypoint motion models and Manhattan against a model of semicircular motion, analyzing the total distance traveled, the average total energy of robots’ agents and convergence time at 80% of exploration of the navigation area using as design methodology, design-based science. Agent's robots involved in the experiment show no cognitive and cooperative parameter, therefore, the results obtained in this article will allow extracting temporal and energy characteristics that allows exploring as much area in the shortest possible time, in order to get heuristics for future work in SLAM tasks