Active Vision for Navigating Unknown Environments: An Evolutionary Robotics Approach for Space Research

Evolved active vision control systems have the ability to extract relevant information from an environment in order to solve a specific task [1]. This work has the objective to investigate an obstacle avoidance and navigation systems for space exploration rovers capable of performing autonomous tasks in challenging planetary terrains. The paper presents an evolutionary robotics approach applied to a Mars rover model that is equipped with an active vision camera and a neural network control system. Preliminary experimental results suggest that such an active vision system provides a powerful and yet computationally cheap way of developing important visual processing strategies to navigate in the environment.