Intercepting a maneuvering target is a challenging problem for a guidance law designer. Recently, a logic-based guidance algorithm (LBA) was developed, assuming that the target performs a randomly switched bang-bang maneuver. The differential game based guidance law, incorporated in the LBA, guarantees hit-to-kill accuracy, assuming perfect information. Since the only direct information, available to the interceptor, is the noisy line-of-sight angle measurements, all state variables should be estimated. In the LBA, the estimators are tuned assuming knowledge of the target switch moments. Thus, the detection of the switch moment becomes the crucial component of the LBA, significantly affecting the interceptor's homing performance. In this presentation, a novel switch detection algorithm, based on a sequentially tested hypothesis, is proposed. The test procedure is derived from an auxiliary convex optimization problem. Algorithm efficiency is demonstrated by examples of numerical simulation.
|Title of host publication
|Proceedings of the 2014 Joint Statistical Meetings (JSM 2014)
|Published - 2014