Abstract

In this work, the performance of two algorithms that allow a mobile robot to guide itself as it explores its surroundings is studied. For that purpose, the performance of harmonic potential fields and A* algorithms is considered. The algorithm offering the best service to a real mobile robot in its future implementation is selected based on test results. Initially, the robot would not know in which environment it is located, and would only obtain limited information about its surroundings by means of its sensors, thus implying that the robot has to trace the route dynamically from the motion starting point to the goal. To evaluate these algorithms, critical surroundings are assumed, which are formed by the possible U- and T-shaped structures that the robot could detect in an unknown environment. The evolution of the dynamic performance of these algorithms generates solutions that represent local minima between the robot and the point objective. The selection criterion of the best algorithm is mainly based on the ability to find the point objective in environments with strong local minima, which need to be solved by means of the studied algorithms to allow the robot to avoid and/or get out of complex structures. Finally, the efficiency of the algorithms in the calculation process, i.e., the time taken by them to reach the point objective satisfactorily, is analyzed and presented, considering that the shorter the time, the greater the selection weighting.