Policy Optimization in Automated Point Merge Trajectory Planning: An Artificial Intelligence-based Approach

Abstract

Air traffic management is a complex decision making process. Air traffic controllers decision on aircraft trajectory control actions directly lead to the efficiency of traffic flow management. This paper aims to realize an automated routine trajectory management in terminal manoeuvring area with an intelligent decision making agent. An artificial intelligence based approach is applied to adaptively and smartly integrate four types of deconflict actions for resolving conflicts. Especially, the reinforcement learning policy optimization process is discussed in detail. Firstly, application of reinforcement learning in adaptive trajectory planning is presented. The entire problem is adaptively divided into several sub-problems. For each sub-problem, an online policy is applied to guide the simulation and optimization modules to find out the conflict free and less delay solution. The online policy is a scale of weight distribution for choosing desirable actions. It follows the rule of roulette wheel selection with weighted probability. The highest desirable decision variable has the largest share of the roulette wheel, while the lowest desirable decision variable has the smallest share of the roulette wheel. Direct policy optimization algorithm is designed to update the online policy. Finally, experiments are built up for validation of the proposed policy optimization algorithm for the intelligent decision making process. The results in the test environment showed that learning agent with different exploration and exploitation ability will result in different system performance in conflict resolution and delay.