提出一种新的多输入多输出(MIMO)非线性检测算法,称为噪声投影按序逐次消去(OSNPC)算法,以改进传统的干扰按序逐次消去(OSIC)算法的计算复杂度方面的性能.OSIC算法从接收信号向量中逐次消去已知的干扰,而OSNPC算法从判决变量向量中逐次消去已知的噪声投影.理论分析表明,OSNPC算法在性能上等效于传统的OSIC算法,但其计算复杂度却大为降低.而且,当OSNPC算法用于对偶格约减(DLR)辅助MIMO检测时,所构成的基于OSNPC的DLR辅助MIMO检测方案,与基于OSIC的DLR辅助MIMO检测方案相比,其整体复杂度得到进一步降低,这是因为在它的检测过程中,省去了对偶格约减基的求逆运算.仿真结果验证了该检测方案的性能与复杂度的理论分析结论. A new multiple input multiple output (MIMO) non-linearity detection algorithm is proposed, which is called the noise projection sequential elimination (OSNPC) algorithm to improve the performance of the traditional interference sequential elimination (OSIC) algorithm in terms of computational complexity The OSIC algorithm successively removes known interference from the received signal vector, while the OSNPC algorithm successively removes known noise projections from the decision variable vector. Theoretical analysis shows that the OSNPC algorithm is equivalent in performance to the traditional OSIC algorithm, The computational complexity is greatly reduced.When the OSNPC algorithm is used to detect the adjacency reduction (DLR) auxiliary MIMO, the OSNPC-based DLR-aided MIMO detection scheme is compared with the OSIC-based DLR-aided MIMO detection scheme , The overall complexity is further reduced, because in its detection process, the elimination of the inverse of the reduction of the dual lattice reduction is avoided.The simulation results verify the theoretical analysis of the performance and complexity of the detection scheme.