Abstract: | We propose a bee-like navigation system for a mobile robot in the complex and natural environment. It has the advantage that object recognition and complicated geometric reasoning are not required. Out of this bee-like navigation system we tackle the problem of estimating vehicle position with respect to a target and reducing the vehicle's errors made during its journey. We achieve this by taking three images with various camera orientations at the position to be memorized. Whenever the robot returns to such a memorized position, comparing current and memorized images using correlation coefficient methods enables the robot to localize itself.
|