【摘要】 针对证据理论无法获得传感器报告、无法处理具有冲突的传感器报告、计算复杂度高、干扰环境下融合结果不可靠等缺点,提出了一种新的辐射源优化识别方法。该方法首先利用灰色关联算法来获得传感器的报告,并且提出利用信息熵解决灰色关联分析中特征权重的选择问题。然后根据传感器证据报告的特点,引入传感器可信度因子,通过构造和分解代价函数将辐射源识别问题转化为求解一个凸二次优化问题。最后,给出了一种利用对数罚函数方法求解该问题的改进方法和步骤。理论分析和仿真结果表明,与证据理论相比,新方法具有更低的计算复杂度、更好的识别能力、更广的适用性和更强的鲁棒性。更多还原

【Abstract】 Aiming at drawbacks of evidence theory,such as it can not get the confidence probability of evidence,and can not deal with high conflicting evidence,and it is high computational complexity,and has uncertainty in interference environment,a new optimization method of radiation source recognition is presented.Firstly,it uses a gray correlation algorithm to get sensor reports,and a new selection method of feature weights based on information entropy is presented.Then,according to characteristics of senor evidence reports,bringing in the reliability of sensor,and by constructing and decomposing cost function,the problem of radiation source recognition is transformed into a convex quadratic optimization problem.Finally,based on the logarithm penalty function method,an improved method and steps are presented.The theoritical analysis and simulation results prove that this new method is less computational complexity,better recognition ability,more widespread availability and stronger robustness that compared with the theory.更多还原