Vision-Based Autonomous Topological Navigation in Outdoor Environments

Abstract

In this paper, we present a proposal of an autonomous topological navigation system for structured outdoor environments. This system uses a camera as the main sensor, and an association of Finite State Machines (FSM) with Artificial Neural Networks (ANN) for path features detection. The environment is represented by a Topological map in which each state is related to a specific track shape. This way, the robot can become able to autonomously reach any destination, going through a sequence of these states, as for example straight path, right and left turns and intersections. The experiments were performed in a real structured outdoor environment in order to validate and evaluate this approach. The proposed system demonstrated to be a promising approach for path features detection.