【摘要】 在分析基于正交频分复用认知系统的局部多窗口联合奇异值分解频谱感知算法优劣的基础上,针对奇异值分解运算量大,对该算法提出了改进。提出了矩阵的F范数代替奇异值分解的方法,以提高算法的实时性。经仿真分析表明,改进后的算法既保证了原有算法中多窗口法(MTM)谱估计的近最优性,提高了算法的实时性,又具有较好的检测性能。更多还原

【Abstract】 With analysis of local multitaper singular-value decomposition-based spectrum sensing algorithm in orthogonal frequency division multiplex cognitive radio system,the algorithm is improved.The norm of the matrix F instead of the singular value decomposition method is proposed,thus to improve the real-time performance of this algorithm.The simulation and analysis shows that the improved algorithm could guarantee the optimality and real-time performance of the original MTM spectral estimation algorithm and also has good detection performance.更多还原