The Study Of Detecting Algorithms In MIMO Systems

Multiple-input multiple-output wireless technology is a great breakthrough in modern digital communications domain. By installing multiple antennas in both of the transmitters and receivers, MIMO system can enhance communication system capacity and spectrum efficiency without increasing bandwidth. MIMO system brings a huge capacity but at the same time it increases the difficulty of signal detection in the receiver, so in the MIMO wireless communication systems the receiver signal detection becomes a key issue.This paper mainly studies the signal detection issues in MIMO systems as follows:First, the paper discusses several detecting algorithms in MIMO wireless communication systems, including maximum likelihood algorithm, the linear detecting algorithm, successive interference cancellation algorithm, ordering successive interference cancellation algorithm, QR decomposition algorithm and sphere detecting algorithm, and compares their BER performance through simulation results.Second, the paper gives a more detailed exposition of the sphere detecting algorithm based on QR decomposition. Sphere detecting algorithm is based on the maximum likelihood criterion produced, it has a reasonable amount of computation and can be achieved the same BER performance of the maximum likelihood detecting. And then introducing sphere detecting algorithm under the searching guidelines of Fincke-Pohst (F-P) and Schnorr-Euchner (S-E) respectively. Then we discuss two pre-processing sphere detecting algorithm and give an comparison in the BER performance and computational complexity with traditional sphere detecting algorithm.Finally, the pre-processing sorting rule of sphere detecting is further studied, a pre-processing sorting rule based on Householder transformation QR decomposition is proposed, and compared with SD algorithm under the rule in both of performance and computational complexity. The superior BER performance and the small computational complexity of heuristic Algorithm are combined in the new pre-processing sorting rule, so the computation of system testing is reduced significantly without sacrificing the BER performance of sphere detecting.