Vision-based Autonomous Navigation Using Neural Networks and Templates in Urban Environments

Jefferson R. SouzaGustavo PessinPatrick Y. ShinzatoFernando Santos OsórioDenis F. Wolf

The aim of this work is to develop a vehicle control system capable of learn behaviors based on examples obtained from human drivers and analyze different levels of memory of the templates (LMT). Our approach is based on image processing, template matching, finite state machine, and template memory. The proposed system allows to train an image segmentation and neural networks which works with LMT in order to identify navigable and non-navigable regions generating as output the steering control and speed for an Electric Autonomous Vehicle, that should stay replicating or improving the human behavior. Several experimental tests have been carried out under different environmental conditions to evaluate the proposed techniques.

Caso o link acima esteja inválido, faça uma busca pelo texto completo na Web: Buscar na Web

Biblioteca Digital Brasileira de Computação - Contato:
     Mantida por: