被动检测系统中自适应算法收敛的动态判断  

Scheme for dynamically judging adaptive algorithm's convergence in passive detection system

作　　者：吴慧娟[1] 文玉梅[1] 李平[1] 

机构地区：[1]重庆大学光电工程学院光电技术及系统教育部重点实验室,重庆400044

出　　处：《仪器仪表学报》2007年第3期414-420,共7页Chinese Journal of Scientific Instrument

基　　金：国家自然科学基金(60374045)资助项目

摘　　要：被动检测应用中,实际环境条件往往比较恶劣而且不可预测,非平稳干扰随时可能发生,如不能准确判断自适应系统实际收敛状况,检测结果将受到严重影响。在实际复杂噪声环境影响下,时间常数已不能代表系统的实际收敛特性,因此提出一种动态判断自适应检测系统迭代算法收敛的方法。基于自适应滤波器工作在理想最优条件时的正交原理,推导出考虑梯度噪声影响时误差与输出信号实际近似正交的约束条件,根据每次迭代不断更新的期望、输出及误差信号的均方估计,在线实时判断收敛并评估自适应解的质量,及时准确停止迭代以得到最佳检测结果,从而避免复杂背景干扰的影响。此动态判断条件对平稳与非平稳环境中的自适应检测均适用,不同条件下的处理结果验证了该方法的有效性。In practical passive detection, the environment settings are always bad and unpredictable, and nonstationary interferences occur from time to time. The detection results will be affected seriously if the convergence of the adaptive system can not be determined accurately. In actual circumstances, time constant no longer represents the practical convergence characteristics of the learning system due to the complicated noise effects. In this paper, based on the principle of orthogonality of the adaptive filter under ideal condition, an on-line constraint of the adaptive detection system, which can be used to dynamically determine the LMS algorithm＇s convergence, is derived when the error and output are approximately orthogonal practically due to the gradient noise. Based on the estimated meansquare values of the mean, output and error signals for each iteration, we can on-line judge the convergence, evaluate quality of the adaptive solution, end the iteration timely and accurately and obtain the optimal detection result while avoiding the complicated background interferences. This dynamic discriminant is applicable to adaptive detection systems both in nonstationary as well as stationary environments. Results processed under various conditions demonstrate the effectiveness of the proposed method.

关 键 词：被动检测 自适应滤波 正交 动态判断收敛 收敛时间 

分 类 号：TN911.23[电子电信—通信与信息系统]