Distributed LOUD detection for true and false warhead recognition

Recognition between true and false ballistic warheads is an important problem in missile defense. In this paper, a multiple-input multiple-output (MIMO) radar with distributed antennas is employed to observe the scene from distinct directions to gain multi-dimensional information of the warheads. A state space model (SSM) is developed to describe the motion behavior of the warheads (state equation) as well as key target parameters obtained by the MIMO radar (observation equation). The parameter matrices (such as the state transition matrix) determined by the motion characteristics of the warheads differ for the true and false warheads, based on which a composite hypothesis test problem is formed for warhead recognition. A distributed detection method based on locally optimal unknown direction (LOUD) test is proposed. We show that, with properly selected system parameters, the detection performance obtained using the proposed method is close to that obtained using centralized processing. The impacts of fusion rule and antenna placement are examined.