Intelligent solutions for cybernetics vehicle control

This work describes a comparison between genetic algorithms (GA) and fuzzy systems paradigms to automatically solve the problem of controlling the vehicle trajectory, enabling to find a target in the Cartesian surface. The fuzzy implementation was based on the work developed by Kosko (1991). In the evaluation of each individual, considering the GA approach genes fitness method was employed ignoring their position in the chromosome. This enables that individuals with different genotypes present the same fenotype, providing a wide genetic diversity. It was confirmed the great flexibility of the GA approach in a continuous changing environment to set the vehicle in the target direction. It was also observed the fast convergence of the GA implementation at the moment that the environment becomes stable: the vehicle keeps its position reaching the target.