提出了一种软件无线电高阶QAM信号的识别算法。该算法直接从接收信号中估计出频偏和相偏并进行补偿,同时估计波特率和定时误差,并经过低通滤波、采样率变换、定时抽取等步骤,恢复出星座图,然后计算星座图实部信息的高阶累积量进行识别。实验结果表明,在SNR大于12dB,码元符号数为1000时,识别率可以达到96%。 A high-order QAM signal recognition algorithm for software radio is proposed. The algorithm estimates the frequency offset and the phase offset directly from the received signal and compensates for them. The baud rate and the timing error are estimated at the same time. After the steps of low-pass filtering, sampling rate conversion and timing extraction, the constellation map is recovered and the constellation is calculated Real part of the information to identify the high-order cumulants. Experimental results show that the recognition rate can reach 96% when the SNR is more than 12dB and the symbol number is 1000.