A generalized oblique projection operator for interference suppression under colored noise

The existing oblique projection (OP) technique usually pays attention to white background noise, but a colored noise assumption is more practical. To solve this problem, a minimized interference constrained generalized oblique projection (MIGOP) operator is proposed under colored noise scenario for interference suppression. It is proven that the OP filter is a special case of the MIGOP filter whereas the latter is more suitable to suppress interference under colored noise scenario. The traditional linearly constrained minimum variance (LCMV) filter and the MIGOP filter are two special cases of the generalized oblique projection (GOP) filters. An LCMV filter can be used to obtain the maximal signal to interference plus noise ratio (SINR), whereas an MIGOP filter can be used to obtain both the maximal SINR and the signal to interference ratio (SIR). The performance comparison among the MIGOP filter, the OP filter and the traditional LCMV filter are illustrated. Conclusions are made based on analysis and simulations.