Pose verification for autonomous equipment interaction in surface mining

It is essential that an automated excavator has accurate knowledge of the position and orientation of a truck that it is to load. The integrity of a given truck pose estimate can be improved by providing the excavator with the capability to verify the accuracy of that estimate using an independent system. A method is presented for verifying the accuracy of estimated truck poses using excavator-mounted range sensors. The method is formulated as a statistical hypothesis test, inspired by Bayesian techniques, and is made computationally feasible with density estimation. Simulation and experimental verification testing shows discrimination between the null and alternative hypotheses, given sufficient measurements of key geometric features of the truck.