The application of GA—Support vector machine in exploration vehicle

A support vector machine (SVM) is applied in the unmanned driving exploration vehicle in this paper. The exploration vehicle is intended for the unknown environment exploration, and will be equipped with six ultrasonic sensors, global position system and a CCD sensor for detecting obstacles. Onboard speed and steering controllers are the core of the guiding system. Measurements of obstacle distance and direction are anticipated to be imprecise however, because the performance of ultrasonic sensors is degraded in complex environments. So we present a support vector machine that can determine a trajectory for an exploration vehicle through unknown environments, even in the presence of imprecise sensor data. The method fully utilizes the potential of the SVM and data fusion to determine vehicle navigation. And the genetic algorithm is used to confirm best parameters of SVM. The simulation results illustrate the robustness of a support vector machine approach regard to sensor imperfections, and could find the optimal path.