基于Jacobi迭代提出一种低复杂度信文检测算法,在算法实现中避免了矩阵求逆运算。数学推导证明,该算法应用于MMSE检测时是收敛的,与传统的Neumann级数展开方法对比,能达到与其完全相同的检测性能,并且在任意迭代次数下能将复杂度保持在2O(K),而后者当级数展开项数i≥3时复杂度上升为3O(K)。为了进一步将Jacobi迭代应用到软判决中,提出了一种用于信道译码的LLR的近似计算方法。仿真结果表明,过几次迭代,Jacobi迭代算法收敛较快并接近MMSE检测性能。 Based on Jacobi iteration, a low complexity message detection algorithm is proposed, which avoids matrix inversion in the algorithm. The mathematical derivation proves that the algorithm converges when applied to MMSE detection. Compared with the traditional Neumann series expansion method, the algorithm can achieve exactly the same detection performance and maintain the complexity at 2O (K) at any iteration. , While the latter increases the complexity to 3O (K) when the number of series expansion i≥3. In order to further apply Jacobi iteration to soft decision, an approximate method of LLR estimation for channel coding is proposed. Simulation results show that Jacobi iterative algorithm converges quickly and approaches MMSE detection performance after several iterations.