This paper establishes the classic linear model of signal of the MIMO radar system with distributed apertures. Based on this model, the design principle and detection performance of MIMO radar detector is investigated under conditions of Gaussian colored noise and partially correlated observation channels. First, the research on design principle of detector shows that the clutter suppression and matched filtering can be independently implemented at each receiving aperture, which greatly reduces the difficulty in implementation of these detectors. Based on these results, a Max detector is proposed for the case where partial channels are disabled due to strong noise and stealth techniques. The second part is the performance analysis of detector. The Fishier divergence coefficient and the statistical equivalent decomposition of limit statistics are used to theoretically analyze the detection performance of AMF detector, and then the analytical expressions of the detection performance of the AMF detector is derived. Analysis results show that both the colored nature of noise and the correlation among observation channels can reduce the capability of spatial diversity of the MIMO radar system, change the target RCSs among observation channels from quick fluctuation to slow fluctuation, and degenerate the detection performance of this radar system into that of the phased array radar system at high signal-to-noise ratio.
