基于压缩感知的谱分析技术是宽带频谱感知的重要方法之一。文章将压缩感知与Welch法相结合,提出了一种新的宽带频谱感知算法。在不重构原信号波形的条件下,直接利用观测矩阵获得的低速观测值进行信号的功率谱估计,降低了采样速率,避免了重构算法的高复杂度。仿真结果表明,与基于压缩感知的多窗口联合奇异值分解算法(CS-MTM-SVD)相比,新算法在压缩率为0.5时能以更高的准确率估计出信号功率谱,估计精度与压缩率的变化成正比关系,并且新算法对噪声容忍度也更高。 Spectral analysis technology based on compressed sensing is one of the important methods for sensing broadband spectrum. The article combines compression perception with Welch’s method, and proposes a new broadband spectrum sensing algorithm. Under the condition of not reconstructing the original signal waveform, the power spectrum of the signal is directly estimated by using the low-speed observation values ​​obtained from the observation matrix to reduce the sampling rate and avoid the high complexity of the reconstruction algorithm. The simulation results show that compared with the CS-MTM-SVD algorithm based on compressed sensing, the new algorithm can estimate the signal power spectrum with higher accuracy when the compression ratio is 0.5, and the estimation accuracy and The change in compression ratio is proportional to the new algorithm and the noise tolerance is higher.