Towards Greener Airport Surface Operations: A Reinforcement Learning Approach for Autonomous Taxiing

抄録

This study proposes an autonomous aircraft taxi-agent that can be used to recommend the pilot the optimal speed profile to achieve optimal fuel burn and to arrive on time at the target position on the taxiway while considering potential interactions with surrounding traffic. The problem is modeled as a control decision problem which is solved by training the agent under a Deep Reinforcement Learning (DRL) mechanism, using Proximal Policy Optimization (PPO) algorithm.1) The reward function is designed to consider the fuel burn, taxi-time, and delay-time. Thus, the trained agent will learn to taxi the aircraft between any pair of locations on the airport surface timely while maintaining safety and efficiency. As the result, in more than 97.8% of the evaluated sessions, the controlled aircraft can reach the target position with the time difference within the range of [-20,5] seconds. Moreover, compared with actual fuel burn, the proposed autonomous taxi-agent demonstrated a reduction of 29.5%, equivalent to the reduction of 13.9 kg of fuel per aircraft. This benefit in fuel burn reduction can complement the emission reductions achieved by solving other sub-problems, such as pushback control and taxi-route assignments to achieve much higher performance.