西安工业信号检测与估计sde_12 ch8 最小二乘估计应用.ppt

* SDE_12 MLE(c) apl * Prony’s Modification to Get Linear Model 用以抵消a参数，先得b，再a * SDE_12 MLE(c) apl * Key Ideas in Prony LS Example 1. Shows power and flexibility of LS approach显示出最小二乘方法的强大能力和适应性 -There is no noise here!!! ?MVU, ML, etc. are not applicable -But, LS works nicely! 没有噪声！但LS照样精确工作 2. Shows a slick trick to convert nonlinear problem to linear one 展示了一种将非线性问题转化为线型问题的技巧 -Be aware that finding such tricks is an art!!!找到那种技巧是艺术 3. Results for LS “Prony”method have links to modeling methods for Random Processes (i.e. AR, MA, ARMA) Is this a practical filter design method? It’s not the best: Remez-Based Method is Used Most * SDE_12 MLE(c) apl * Ex. 8.13 Adaptive Noise Cancellation 自适应噪声对消器 抑制背景噪声 1：工频干扰 2：胎儿心电监护中抑制母亲心跳 与i统计相关但与d[n]统计无关 参考通道 希望获得的 干扰 去除干扰的信号 估计干扰的最佳对消信号 * SDE_12 MLE(c) apl * Noise Cancellation Typical Applications Fetal Heartbeat Monitoring胎儿心电监护仪 胸膛 腹部 * SDE_12 MLE(c) apl * 2. Noise Canceling Headphones * SDE_12 MLE(c) apl * 3. Bistatic Radar System收发分置雷达 * SDE_12 MLE(c) apl * LS and Adaptive Noise Cancellation 衰退记忆 遗忘因子 干扰非平稳 * SDE_12 MLE(c) apl * 具有遗忘因子的序贯Sequential LS with Forgetting Factor We want to weight recent measurements more heavily than past measurements… that is we want to “forget” past values.噪声在变，所以现在比以前权值重 So we can use weighted LS… and if we choose our weighting factor as an exponential function then it is easy to implement! * SDE_12 MLE(c) apl * 综合示例：TDOA方位估计 Ex. 6.3 基于到达时间差的源定位 TDOA-based Emitter Location 假设信号像脉冲，可通过上升沿测量时间t this assume signals are pulse-like so you can estimate by locating the rising edge to measure t: TDOA Time Difference of Arrival 到达时间差 TOA Time of Arrival 到达时间 假定3个接收器放在已知的位置， 不知道信号发出的时间t， 仅知道信号的到达时间t1 t2 t3 t1 t2 t3 Hyperbola t1-t2 Hyperbola t2-t3 由时间差 t1－t2可以确定位于双曲线Hyperbola 1 由时间差 t2－t3 可以确定位于双曲线Hyperbola 2 由曲线交汇点可以确定源位置 炸点定位 TDOA 原理 * SDE_12 MLE(c) apl * t1-t2：计算？ 炸点—好说，但对于一般信号：？？？？ n0 * SDE_12 MLE(c) apl * Emitter Location is Two Estimation Problems in One: 1) Estimate Signal Parameter(s)