Blind DOA estimation using higher order cyclostationarity and linear prediction in multipath environment

In this paper, we propose a new approach that combines cyclostaliomirity and linear prediction (LP) to extract the cyclic correlations of only signals with the same cycle frequency and out the cyclic correlations of stationary additive noise and all other interferences with different cycle frequencies in coherent environment. Unlike similar existing methods, the proposed algorithm requires only a uniform linear subarray and the rest of array may have arbitrary and unknown response. The maximum number of signals resolvable with our method may exceed the number of sensors. Simulations are conducted to prove the effectiveness of the proposed approach.