Template-based autonomous navigation and obstacle avoidance in urban environments

Abstract

Autonomous navigation is a fundamental task in mobile robotics. In the last years, several approaches have been addressing the autonomous navigation in outdoor environments. Lately it has also been extended to robotic vehicles in urban environments. This paper presents a vehicle control system capable of learning behaviors based on examples from human driver and analyzing different levels of memory of the templates, which are an important capability to autonomous vehicle drive. Our approach is based on image processing, template matching classification, finite state machine, and template memory. The proposed system allows training an image segmentation algorithm and a neural network to work with levels of memory of the templates in order to identify navigable and non-navigable regions. As an output, it generates the steering control and speed for the Intelligent Robotic Car for Autonomous Navigation (CaRINA). Several experimental tests have been carried out under different environmental conditions to evaluate the proposed techniques.