engleski

Hybridization of adaptive genetic algorithm and ART 1 neural architecture for efficient path planning of a mobile robot

This paper presents a novel and efficient Knowledge Based Genetic Algorithm (KGA) developed for finding feasible paths in an environment containing arbitrarily distributed obstacles. Furthermore, the neural architecture is incorporated into the robot controller to enable the incremental learning of spatial distribution of obstacles and the formation of experience. Combining these two techniques enables the simulated robot to efficiently navigate through the space, to continuously increase the level of knowledge, and to reuse the experience formed by exploration. The approach is tested on a series of simulated robots and environments.

path planning; genetic algorithm; mobile robot; optimization; neural network

nije evidentirano