分布式探测是雷达领域热点问题,信号级融合探测比数据级融合探测能力更强,但通常需要的通信带宽较大。为此文中针对分布式非相参信号级目标融合探测,提出了基于雷达压缩数据的信号级融合目标检测方法。所提方法通过可并行化计算的信号级融合算法实现不同雷达量测值之间的去耦,通过双门限检测避免传输局部低能量的噪声信号,通过二次量化对过门限信号进行再次压缩,最终实现以点迹通信带宽逼近信号级融合检测的能力。基于4 雷达组网的数值仿真结果验证表明,通信带宽缩减至原来的1/1 000,信噪比损失不超过0. 7 dB,并据此探索雷达组网的体系架构设计问题,可支撑不同场合下的信号级协同探测工程应用。

Distributed detection is a hot topic in the radar field. Signal fusion-based detection generally outperforms data fusion based detection, but the communication cost is often huge. In order to tackle this problem, a data compression algorithm for distributed non-coherent target detection based on signal fusion is presented in this paper. In the proposed algorithm, signal fusion with parallelized computation is employed to realize the decoupling of observations from different radars, and censored detection is used to eliminate locally unpowerful noise from transmitting, and then censored observations are compressed by requantization processing. Detection performance of the proposed algorithm is capable of approaching a signal fusion-based algorithm, but only needs a low communication cost like a data fusion-based detection. Numerical simulation results with four distributed radars indicate that compared with signal-fusion based detection algorithms, the communication bandwidth of the proposed compression algorithm can be reduced to 0. 1??, whereas the signal-to-noise ratio loss is less than 0. 7 dB. Accordingly, the radar network structure design problem is then discussed for distributed radar to support different application scenarios.

TN957