Symbol decision feedback equalization based joint detection algorithm

The traditional joint detections use the block linear equalization and block decision feedback equalization,They all handle the equalization and detection with a whole data block,so the system matrix is much larger.In this paper the system matrix is divided into the minimal detectable units,that are data symbols,to jointly detect data symbol by symbol and finally generate a symbol linear equalization based joint detection.Neitherless we apply the principle of block decision feedback equalization into it to create a symbol decision feedback equalization based joint detection.These algorithms have much smaller system matrix,so the computation complexities are much lower.The symbol decision feedback equalization can reduce the noise item and the previous symbol decision error item,compared to the symbol linear equalization,so the overall performance approaches to the block decision feedback equalization.The experiments approve it.We also discuss the algorithm modification in case that different users are allocated with different spreading factor,and propose an improvement to sort the users data symbols by the magnitutes of the users spreading factor.The experiment approves the advantage of this improvement.