Channel estimation in MIMO feeder links using superimposed test signals

To achieve optimum operation, one needs to continuously estimate the channel coefficients of MIMO feeder links and adapt their precoding matrix accordingly. A technique is proposed to achieve channel estimation in a 2x2 MIMO feeder link scenario consisting of superimposing test signals to the data streams at the ground stations. At the receiving end, sequences of the symbols comprising the test signals are stored and through correlation with the received signals one should be able to estimate the coefficients of the channel. It is essential that the amplitude of the test signals is much smaller than that of the data as the superposition causes a small degradation, which should be kept as small as possible.
The available literature has shown that CAZAC sequences are the best candidate as test signals to fulfill the channel estimation performance requirements. However, the procedure is impaired by phase scintillations in the atmosphere and is highly dependent on the characteristics of the CAZAC sequences. Hence, it is of high interest to have a simulation environment in MATLAB for the proposed technique. It should consider the effect of atmosphere perturbation and the effect of using different CAZAC sequence lengths. An initial model is already available, and it is expected that the work starts from there.