在电子对抗领域,信号调制方式识别是进行雷达分选、干扰施放的基础,得到广泛研究。对此,文中提出了一种以信号频谱相像系数和幅度统计参数为分类特征的FSK/BPSK复合调制雷达脉冲信号识别算法。算法首先提取雷达脉冲信号的频谱相像系数和幅度统计参数,然后采用分层结构的神经网络分类器进行识别。该算法不仅能识别FSK/BPSK复合调制信号,且对其他常用雷达信号调制方式的识别不产生干扰。仿真结果表明,针对FSK/BPSK以及CW、LFM、BPSK、QPSK、FSK等常用雷达信号调制类型,在信噪比>5 d B时,分类正确率可达98%以上。
For the frequency difference of arrival(FDOA) estimation in passive location, this paper transforms the frequency difference estimation into the radial velocity difference estimation,which is difficult to achieve a high accuracy due to the mismatch between the sampling period and the pulse repetition interval. The proposed algorithm firstly estimates the point-in-time that each pulse arrives at two receivers accurately. Secondly two time of arrival(TOA) sequences are subtracted. And finally the radial velocity difference of a target relative to two stations with the least square method is estimated. This algorithm only needs accurate estimation of the time delay between pulses and is not influenced by parameters such as frequency and modulation mode. It avoids transmitting a large amount of data between two stations in real time. Simulation results corroborate that the performance is better than the arithmetic average of the Cramer-Rao lower bound(CRLB) for monopulse under suitable conditions.
