Adaptive blind diversity combining for antenna arrays

Adaptive antenna array is a powerful technique used for maximising directional gain. It reduces effectively the outcome of fading and interference caused by multipath propagation in dispersive channels. One particular case of this technique is known as equal gain diversity combining where signals from two or more antennas are combined in phase. This process often needs an initial training sequence necessary for the combiner to match the phases of the signals arriving from the antenna branches. However, changes in the communication channel require periodical updating of the phase combiner and restart of the training process. In this work we present a diversity combining method that does not need to use training sequences for computing the antenna gains, for signals modulated with constant modulus envelope (e.g. M-ary PSK, MSK, FM, etc.). The method relies on statistics of higher order of the incoming signal, and the proposed algorithm can be classified as blind deconvolution. It presents a good tracking capability for the optimum phase combination of the antenna branches. The system is designed so as to maximise the signal-to-interference-plus noise power ratio (SINR) at the output of the array. This is analogous to the process of adaptive beam forming in which an array antenna modifies its radiation pattern to increase the incoming wanted signal power. Simulation results are presented as a complement to the theoretical analysis