Arthur Lue, MIT Lincoln Laboratory
Keywords: Astrodynamics
Abstract:
A critical problem is emerging in the space environment. For the foreseeable future, we expect there to be a collision between a large object and a debris object every 400 days. But these collisions are hidden amongst 10,000s of close conjunctions per day and pose catastrophic threats to mission-critical payloads. This report describes work addressing the problem of systematically identifying these needles in the proverbial haystack. Several key avenues are pursued: developing a computational scheme that can rapidly estimate conjunctions for large catalogs (~100,000 objects); improving covariance analysis to effectively cull the number of encounters of critical interest; devising new covariance-based statistics such that, regardless of orbit-estimation quality, all true conjunctions of critical interest are successfully tagged; introducing an automated, adaptive tasking scheme to ascertain all potential collisions with a low false alarm rate; and providing quantification for tying conjunction-prediction performance to sensor-network performance and tasking parameters. As a specific example, by taking a 50,000 object catalog in the most dangerous debris regime, we established a protocol where sensors take a series of observations over the course of 10 days to predict the 1.4 conjunctions of less than 10 m for the 11th day. Only 9 conjunctions per day need to be monitored to identify those 1.4 true conjunctions. The sensor-network load necessary is 3.8 times the loading necessary to maintain the latent day-to-day catalog.
Date of Conference: September 13-16, 2011
Track: Astrodynamics