Linear prediction based semi-blind estimation of MIMO FIR channels

The multichannel aspect has led to the development of a wealth of blind channel estimation techniques over the last decade. However, most of these blind techniques are not very robust and only allow us to estimate the channel up to a number of ambiguities, especially in the MIMO case. On the other hand, all current standardized communication systems employ some form of known inputs to allow channel estimation. The channel estimation performance in those cases can always be improved by a semi-blind approach which exploits both training and blind information. The purpose of this paper is to introduce semi-blind techniques of which the complexity is not immensely much higher than that of training-based techniques. The semi-blind criteria are quadratic and combine a training-based least-squares criterion with a blind criterion based on linear prediction. A variety of convenient linear prediction approaches are considered.