Robotic Behavior Implementation Using Two Different Differential Evolution Variants

In Evolutionary Robotics, Bioinspired Algorithms are used to generate robotic behavior. Several researchers used classic Genetic Algorithms or adaptations of Genetic Algorithms for developing experiments in ER. Here, we use Differential Evolution as an evolutionary alternative method to automatically obtain robotic behaviors, selecting a wall-following behavior as a representative example. We used an e-puck robot and the Player-Stage simulator for the experiments. We detail the results and the advantages when using the DE variants in our application with the simulated and the real robot. In order to optimize time for evolution, and test the resultant behavior in the e-puck robot, for our experiments we employed the Player-Stage simulator.