Invariant tests for spatial stationarity using covariance structure

The authors exploit the structure of covariance matrices of stationary processes to obtain invariant tests for stationarity, especially spatial stationarity. With this objective, they characterize the maximal invariants for testing for Toeplitz and centrosymmetric structure. This approach seems new and leads to invariant CFAR tests. The distribution is obtained under the null hypothesis and under the alternate hypothesis for a uniform linear array with a single near field interferer and an arbitrary far field environment. Simulation results indicate excellent performance.<>