Trajectory Planning for Conflict-free Trajectories: A Multi Agent Reinforcement Learning Approach

While data-driven methods aim to build models for trajectory planning and conflicts resolution, incorporating stakeholders’ interests and preferences, the multi-agent reinforcement learning (MARL) approach aims to address complexity phenomena due to traffic and resolve conflicts between multiple trajectories, simultaneously. Towards that goal we aim to formulate the problem as a Markov Decision Making Process, i.e. MDP, and apply multi-agent reinforcement learning (MARL) methods to resolve it.

Contributions expected:
1. Contribute towards collaborative decision making for conflict-free trajectory planning involving multiple stakeholders (Airspace users, ANSPs, Network Manager, Traffic Controllers).
2. Reveal preferences and constraints concerning trajectories and airspace use using data-driven approaches.
3. Explore and provide an understanding of how resolution of conflicts among trajectories is affected by preferences and constraints concerning trajectories and airspace use.
4. Reduce Air Traffic Controllers workload due to conflicts at the tactical stage, while enabling performing effective conflict-free planning at the planning stage of operations.
5. Explore and develop where appropriate spatiotemporal abstractions in MARL techniques to address large-scale complex phenomena in real-world settings.
6. Contributions to the ATM Master Plan:
 Improved operations productivity via contributions to improved collaborative planning tools accounting for complex phenomena due to traffic.
 Increasing predictability via efficient operation plans devised at the pre-tactical stage, reducing buffers and uncertainty.
 Minimizing the impact of delays due to ATM factors.
 Reduce number of flight inefficiencies due to tactical ATC actions, supporting better planning of operations for Airspace Users.