在不同探测环境下确定探测概率时 ,首先需要根据给定的虚警率来确定相应的探测概率门限因子 ,这时需要在杂波和被检测单元的概率分布密度函数已知条件下 ,对其进行积分 .实际应用中 ,概率分布密度函数非常复杂 ,这时直接积分法很难奏效 ,必须用数值模拟法来解决 .但常规的 Monte Carlo方法在模拟像虚警概率这类小概率事件时既耗费机时 ,精度又差 ,目前国际上正在研究应用重要度抽样法(Importance Sampling)来解决该问题 .本文在前人研究成果的基础上 ,针对两大类雷达探测处理器——单元平均恒虚警 (CA-CFAR)和修整平均恒虚警处理器 (TM-CFAR) ,对相应的重要度抽样方法进行了研究和推导 ,结果表明方法高效可靠.

During the performance evaluation of radar system, the threshold multiplier should be determined firstly for a given system probability of false alarm. For this purpose, we must know the clutter probability distribution of various CFAR scheme. In most departures from the Cell Averaging constant false alarm rate (CA-CFAR) structure, serious difficulties arise in performance analysis for various clutter distribution. Importance sampling (IS) method is a weighted Monte Carlo method for rare event simulation. ...