Improved Track Accuracy for Missile Engagements

In this proposal, Reservoir Labs introduces an innovative approach to the problem of orbital track reconstruction and discrimination from radar data. We propose a technique that utilizes radar returns over a much longer observation window than is current (minutes and tens of minutes not seconds), for greatly improved integrated tracking and discrimination. As an example, we extract key information related to ballistic missile break-up events by working in a state-space of orbital invariants only. The key to our approach is the space-time viewpoint, where time is not measured by the counting of measurements, or by seconds on a clock, but by the accumulation of information from measurements. This line of thinking leads not to a conventional track filter, but to a set of static constrained optimization problems that we attack using new methods arising from research in the field of convex optimization. Clutter and debris are very real practical challenges to tracking in the types of environments encountered in BMD. This challenge is also addressed head-on, by introducing a detection-level filtering concept for debris. During Phase II of this program, we will focus on demonstrating our algorithms on measured data and creating a software prototype for test and evaluation. Approved for Public Release | 16-MDA-8951 (15 December 16)