Inertial Navigation Compensation with Reinforcement Learning

This paper presents a method for applying Reinforcement Learning (RL) techniques to extend the holdover time of an inertial system in the absence of aiding from a Global Navigation Satellite System (GNSS). Several RL algorithms were evaluated using this method. The performance results, in terms of positional error, for each algorithm are compared to each other as well as to the results from an unaided Kalman Filter and a navigation-grade Inertial Navigation System.