基于连续信任函数理论的鲁棒雷达目标检测方法

目标检测是雷达的基本功能, 其鲁棒性直接决定了雷达应用的成败。针对瑞利目标检测问题, 充分利用连续信任函数理论在不确定性信息表征和推理方面的优势, 提出一种鲁棒雷达目标检测方法, 在决策空间中可对先验模型噪声、真实目标和外来干扰的信任进行精准量化。鉴于该检测方法不存在解析解的问题, 通过随机变量的变换, 将任意瑞利目标幅度起伏模型转化为标准瑞利模型, 再利用查表方式实现信任函数的数值计算, 最后完成检测决策或用于后续融合推理。瑞利目标检测实验验证了所提方法的鲁棒性。

Robust radar target detection method based on continuous belief function theory

Target detection is the basic function of radar, and its robustness directly determines the success or failure of radar application. Aiming at the problem of Rayleigh target detection, a robust radar target detection method is proposed by making full use of the advantages of continuous belief function theory in uncertain information representation and reasoning. The belief of prior model noise, real target and external interference can be quantified accurately in the detection decision space. In view of the fact that the proposed method does not have analytical solution, through the transformation of random variables, the arbitrary Rayleigh target amplitude fluctuation model is transformed into the standard Rayleigh model.And then, the numerical calculation of the belief function is realized by using the look-up table method. Finally, the detection decision is completed or used for subsequent fusion reasoning. Rayleigh target detection experiments verify the robustness of the proposed method.