Exploiting the unexpected: Negative evidence modeling and proprioceptive motion modeling for improved Markov localization

Abstract

This paper explores how sensor and motion modeling can be improved to better Markov localization by exploiting deviations from expected sensor readings. Proprioception is achieved by monitoring target and actual motions of robot joints. This provides information about whether or not an action was executed as desired, yielding a quality measure of the current odometry. Odometry is usually extremely prone to errors for legged robots, especially in dynamic environments where collisions are often unavoidable, due to the many degrees of freedom of the robot and the numerous possibilities of motion hindrance. A quality measure helps differentiate the periods of unhindered motion from periods where robot motion was impaired for whatever reason. Negative evidence is collected when a robot fails to detect a landmark that it expects to see. Therefore the gaze direction of the camera has to be modeled accordingly. This enables the robot to localize where it could not when only using landmarks. In the general localization task, the probability distribution converges more quickly when negative information is taken into account. © Springer-Verlag Berlin Heidelberg 2006.