In this talk, we will present our research on cooperative optimal defensive strategies in multiple-vehicle engagements. Specifically, we consider the scenario in which two target aircraft are being pursued by two missiles. A novel defensive guidance law will be presented in which the target pair maneuvers cooperatively to lure their pursuers into collision. Unlike defensive strategies such as evasion or deploying defender missiles, the proposed guidance law does not require the targets to have acceleration advantage over their pursuers nor do they need to carry additional missiles.
The talk will be split into two sections. In the first section, derivation of the cooperative guidance law based on linear quadratic optimal control theory will be presented. The derivation of the optimal cooperative strategy assumes that the pursuing missiles are using linear guidance laws and all vehicles have constant speeds with arbitrary-order linear dynamics. We will demonstrate the performance of the cooperative strategy when using perfect state information through a “collision map”. This map highlights the ability of the proposed guidance law to lure the missiles into collision over a range of initial conditions. In the second part of the talk, we will consider the case when the targets have imperfect information on the missiles’ states and guidance laws. Here we will present a decentralized estimation/ identification scheme that is based on the multiple model adaptive estimation (MMAE) approach. However, unlike the conventional MMAE approach in which the number of required filters for guidance law identification grows quadratically with the number of possible missile guidance strategies, number of filters in the proposed scheme only grows linearly and is computationally more efficient. Missile-missile miss distance performance of the cooperative guidance law with estimator-in-the-loop will be demonstrated through a Monte Carlo simulation study in which we will also highlight its sensitivity to missile-target range estimation error.
