Biologically inspired vision based control using featureless time-to-contact estimations

Time-to-contact, a biologically inspired concept, is defined as the time for a moving observer to contact an object or surface if the current velocity is maintained. Since it can be directly estimated from image sequences using a vision sensor, time-to-contact is widely applied for various robotic applications such as docking, chasing, landing, and navigation. Nevertheless, existing estimation methods, such as the size based method and optical flow divergence method, require the time-consuming feature extraction and tracking. In this paper, we adopt a featureless method that can directly estimate time-to-contact using only the image intensity information. Moreover, we propose an error based proportional control algorithm based on the estimated time-to-contact. Such a controller is implemented on a mobile robot platform to achieve real-time docking. We have also proposed a new time-to-contact estimation algorithm considering the robot's angular velocities and verified the result using image sequences. The estimation and control method presented in this paper has the potential to be applied to miniature robots to accomplish agile locomotion with limited computation power or limited sensing modalities.