A neuroinspired cognitive behavioral control architecture for visually driven mobile robotics

Abstract

Several studies have shown that the optic flow serves as a tool for navigation for animals. Flying insects use it to follow paths and avoid obstacles, while in primates it represents an additional input that can improve navigational performance. A neuroinspired architecture for optic flow calculation and decision making, based on the cortical organization of the human brain, has been developed and successfully used as a novel control system for a mobile robot. Navigation in a corridor and obstacle avoidance are achieved relying only on the optic flow. Moreover, the presence of the optic flow improves the performance of the robot in target reaching and obstacle avoidance tasks in a virtual environment, affecting its trajectories as shown by experiments with human subjects.