This repository contains an implementation of the conflict resolution portion of the UAS Traffic Management project using Discra (Distributed Conflict Resolution Architecture). While this work uses ideas developed in the paper titled "Short-Term Conflict Avoidance for Unmanned Aircraft Traffic Management" by Hao Yi Ong and Mykel J. Kochenderfer (see our paper), the focus here is to create an efficient system to handle a large volume of conflict resolution requests in a distributed streaming setting.
This implementation is meant to accompany an upcoming paper under review for an AIAA journal and is completely local. It does not include code that communicates with the UTM client server. If that is the intent of the user, please contact the authors.
Ensuring safety and providing timely conflict alerts to small unmanned aircraft is important to their integration into civil airspace. This prototype software provides advisories, when necessary, for short-term conflict avoidance as part of an automated traffic management system. The goal is to balance aircraft safety and efficiency subject to environmental and aircraft uncertainty. The controller generates advisories for each aircraft to follow, and is based on decomposing a large Markov decision process and fusing their solutions. We separate the problem into pairwise encounters that are solved offline. The solutions to these encounters are then combined online to produce a locally optimal solution using an iterative search technique. As a result, the method scales well and solves the global problem efficiently.
Detailed documentation for Discra and instructions on how to modify the conflict resolution algorithm implemented in the architecture can be found here.
If you use Discra for published work, we encourage you to cite the software using the following BibTex citation:
@InProceedings{ong2015,
Title = {Short-term conflict resolution for unmanned aircraft traffic management},
Author = {Ong, Hao Yi and Kochenderfer, Mykel J.},
Booktitle = {IEEE/AIAA Digital Avionics Systems Conference},
Year = {2015}
}