可并行化计算雷达网信号级融合检测算法研究

In the distributed radar research, cooperate detection based on signal fusion is a critical theoretical issue for the next generation netted radar system. However, signal fusion based detection faces many tough problems, such as heavy communication and computation demand, scheme difference of relevant radars and signal-noise ratio difference. This project will study the signal fusion based detection algorithms for distributed radar and will try to solve all the problems by studying fusion algorithms with the parallel computable capability. Such algorithms can decompose the heavy computation burden for calculating the final test statistic into distributed radars, such that each radar node can use its computation resources and transmit only the results to the signal fusion center for the final decision. Parallel computable signal fusion based algorithms can realize signal fusion based detection at a low cost and enable flexible netting for signal fusion based detection. This research may accelerate the engineering application of the radar network with the signal fusion based detection.

多站雷达协同信号级融合检测是下一代雷达系统的关键理论问题，但实现信号级融合检测面临通讯量大、计算量、雷达体制异构、信噪比不同等问题。本项目将对多站雷达基于信号级融合的检测目标检测算法进行研究，试图通过研究具有可并行化计算能力的信号级融合目标检测算法的设计同时解决以上问题，这种信号级融合检测算法的检验统计量的计算可被并行化的分解各自雷达站中，各雷达站可利用自身的计算资源并只需将计算结果传输至信号融合中心进行最终的融合判决。可并行化计算信号级融合检测算法可以较低代价实现信号级融合的检测性能，基于该算法的信号级融合网络组网方式灵活，可为信号级融合雷达组网系统的工程应用提供良好的工程支撑。

多站雷达协同信号级融合检测是下一代雷达系统的关键理论问题，但实现信号级融合检测面临通讯量大、计算量、雷达体制异构、信噪比不同等问题。本项目将对多站雷达基于信号级融合的检测目标检测算法进行研究，试图通过研究具有可并行化计算能力的信号级融合目标检测算法的设计同时解决以上问题，以标准信号融合、双门限融合检测和二次量化融合检测为准则，研究了单站基于平方律检波，单元平均恒虚警和广义似然比的融合检测算法，它的计算量被并行化的分解各自雷达站中，只需将计算结果传输至信号融合中心进行最终的融合判决，可以较低代价实现信号级融合的检测性能，可为信号级融合雷达组网系统的工程应用提供良好的工程支撑。