Vision Based Navigation System of Autonomous Mobile Robot. Planning of Landmark Sensing Considering Environmental Conditions.

Abstract

We propose the Planning of Landmark Sensing (PLAS) considering environmental conditions for autonomous mobile robots.When a robot moves, the robot usually resets accumulated position errors by sensing the landmark.In previous works, these errors are estimated based on robot models, and Kalman Filter is used to reset the errors after the landmark sensing.Based on this method, only the relation of the robot and the landmark position is considered to influence the observation noise of Kalman Filter.But in the actual visual landmark sensing, environmental conditions, like position of lighting and brightness of room, are likely to cause the misrecognition of landmark.Then we consider environmental conditions for the observation noise.This helps our system to recognize the possibility of the misrecognition in bad sensing environments.As a result of the consideration of the environmental conditions, our navigation system can navigate robots precisely, even if the environments are not suitable for the landmark sesning.