Robot Depth Estimation Inspired by Fixational Movements

Distance estimation is a challenge for robots, human beings, and other animals in their adaptation to changing environments. Different approaches have been proposed tackle this problem based on classical vision algorithms or, more recently, deep learning. We present novel approach inspired by mechanisms involved fixational movements estimate depth image with monocular camera. An algorithm microsaccades head during visual fixation presented. It combines the images generated these micro-movements ego-motion signal, compute map. Systematic experiments using Baxter robot Gazebo/ROS simulator are described test two different scenarios evaluate influence of its parameters robustness presence noise.