An eigenvalue decomposition and interference cancellation array processing algorithm for low mobility CDMA communications

We present a new beamforming algorithm which is based on eigenvalue decomposition and interference cancellation (EDIC) for application in low mobility DS/CDMA communication systems. Using the desired signature code and an orthogonal to it code, we can decompose the receiving signals into two subspaces: signal and interference. The EDIC beamformer can handle a large number of users with a small number of antenna elements and the weights can be blindly updated according to these two subspaces. A recursive algorithm with low computational complexity has been developed for IS-95 systems. In slow fading environments, we further reduce the computational complexity by down sampling the interference and by slowly tracking the desired channel response. Simulation results show that the new recursive algorithm reduces the computational complexity by 98% with a performance drop of 2 dB.