A robust Bayesian agent for optimal collision avoidance manoeuvre planning

Greco, C. and Sánchez, L. and Manzi, M. and Vasile, M. (2021) A robust Bayesian agent for optimal collision avoidance manoeuvre planning. In: 8th European Conference on Space Debris, 2021-04-20 - 2021-04-23, ESA/ESOC.

[thumbnail of Greco-etal-ECSD-2021-A-robust-Bayesian-agent-for-optimal-collision-avoidance]
Text. Filename: Greco_etal_ECSD_2021_A_robust_Bayesian_agent_for_optimal_collision_avoidance.pdf
Accepted Author Manuscript
License: Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 logo

Download (1MB)| Preview


This paper addresses the problem of automatically allocating Collision Avoidance Manoeuvres under uncertainty by a robust Bayesian framework. This framework allows propagating the objects' uncertainty, predicting collisions, allocating manoeuvres, updating the state estimation with Bayesian inference, and redefining the manoeuvres, accounting at all steps for aleatory and epistemic uncertainty. The Bayesian framework combines a robust particle filter for state estimation and uncertainty propagation, an intelligent agent for automatically classifying risk events and allocating avoidance manoeuvres, and a Collision Avoidance Manoeuvre optimiser for obtaining the optimal manoeuvre. A test case is included to show the operation of the system. Two scenarios are presented: a collision and a near-miss conjunction.