贝叶斯检测跟踪联合处理的检测阈值设置方法

基于贝叶斯理论的检测跟踪联合处理方法能够利用目标的运动模型在帧数据间以概率的形式对信号进行积累，从而提高系统的检测性能。然而，在Neyman-Pearson准则下，该方法很难根据系统要求的虚警概率计算检测阈值，影响了该方法的应用范围。在综合分析基于贝叶斯理论的检测跟踪联合处理方法的基础上，针对Neyman-Pearson准则，详细推导了系统虚警概率同检测阈值之间的关系，并在观测噪声为高斯白噪声的情况下给出了检测阈值的近似闭式解。利用该闭式解，可以按照系统要求的虚警概率计算检测阈值，从而使得实际系统的虚警概率满足要求。最后，计算机仿真实验验证了该闭式解的有效性。

New approach to determine detection threshold for Bayesian track-before-detect

Bayesian track-before-detect can integrate signals in the form of probability among data frames by making use of target motion model,which can improve the detection performance for stealthy targets.However,the detect threshold of Bayesian track-before-detection is difficultly set under Neyman-Pearson criterion.Based on the analysis of basic principles of Bayesian track-before-detect,the paper analyzes the relationship between the false alarm probability and the detection threshold in detail,and presents an approach to determine the detection threshold in white Gaussian noise.Using the approach,the detection threshold can be ascertained according to the false alarm probability in demand.Finally,the computer simulation experiments verify the feasibility of the approach of determining the detection threshold.