Transmitters Selection in Distributed MIMO Radar Network with Partially Correlated Signals

In this paper, we consider a problem of transmitters selection for parameter estimation in a distributed multi-input multi-output (MIMO) radar network. Waveforms of the transmitters in the radar network are assumed to be partially correlated. The estimation performance is characterized by the inverse of Fisher information matrix (FIM) as Cramér Rao lower bound. Different from those transmitters with orthogonal waveforms, the non-orthogonality of probing signals leads to the non-diagonal structure of the FIM. Owing to this fact, the seeking of optimal subset of transmitters is formulated as a sensor selection problem with correlated measurement noise. To efficiently address nonconvex integer programming problem, we propose and verify a novel equivalent relaxation. Using the bi-convexity of proposed surrogate problem, the suboptimal solution is numerically calculated by a algorithm based on nonlinear alternating direction method of multipliers. Simulation results are presented to assess the effectiveness and reveal the superiority of the proposed method.