【作者】 曹轶超；

【导师】 顾国祥； 方建安；

【作者基本信息】 东华大学 ， 控制理论与控制工程， 2010， 博士

【摘要】 无线定位技术是备受各方关注的实用技术,也是移动机器人、无线传感器网络等前沿研究中的关键技术。其中与定位算法有关的信号处理技术成为直接影响定位精度、算法稳定性和实时实现的核心所在。尽管在该领域的学术研究不断深入,但仍然有许多技术瓶颈和实际问题需要研究解决。尤其是在无人机和无线传感器网络等低成本、超小型、网络化探测形式广泛运用后,相应的适用技术研究滞后,实用技术优化方案缺乏。本论文针对高斯噪声环境下静、动态点目标的无线定位、追踪与运动分析等信号处理问题进行了算法研究,着眼于提高精度与鲁棒性、低成本解决方案、实用化技术等方面,为发展低成本的实用性目标运动分析方法探索可行的途径。主要的研究内容和创新性研究成果如下:(1)针对静态目标的无约束定位算法研究。系统地讨论了基于波达角、波达时间差的静态目标定位估计算法,结合与无线定位相关的非线性统计估计理论,计算了基于TDOA、AOA、TDOA/AOA无约束优化定位算法的Fisher信息阵。通过引入类线性模型,论证了模型中的噪声项在宽松的条件下接近高斯分布,得到了针对问题的近似极大似然算法。仿真实验显示近似ML算法较LS算法更优越。(2)基于TDOA的约束WLS无线定位算法研究。提出了用带二次型约束的类线性模型来准确描述TDOA无源定位问题,并用同时对角化技术辅助,给出了二次型约束加权最小二乘WLS解。在假设适度信噪比的高斯噪声环境下,新算法是一种近似极大似然算法。仿真实验显示新算法高效且比已有算法在定位精度方面有显著提高。(3)基于约束Kalman滤波的定位与跟踪方法研究。研究了基于TDOA和DOA的运动目标定位与追踪问题。鉴于地面移动目标状态向量的随机性,提出用约束的数学期望来代替硬二次型约束,并推导了带约束的线性最小均方误差估计器。提出用带二次型约束的伪线性模型描述状态向量的动态特性,并用约束Kalman滤波来解决基于连续TDOA和DOA测量的目标定位和跟踪问题,把非线性项作为一个状态变量来估计。仿真实验显示具有简单递归形式的约束Kalman滤波算法能有效地追踪运动目标,计算复杂度较传统Kalman滤波算法增加很少。(4)基于功率和多普勒频率的目标运动分析算法研究。结合发展适用技术的理念,探讨了适合在简单超小的无线电测量装置上实现的目标运动分析算法,该算法可用于在噪声环境下追踪移动无线电信号源。分析显示功率测量与多普勒频率测量的Fisher信息矩阵具有互补性。分别对基于接收功率和频率的目标运动分析问题给出了迭代最小二乘解法,并通过仿真实验对比了基于功率、频率及两者结合算法的估计误差,总结出针对问题的最佳适用技术方案。(5)基于传感器网络测量峰值功率的目标运动分析算法研究。延续发展适用技术的理念,构造了一种低成本、高效的传感器网络对发出电磁波的移动目标进行探测的机制。提出了传感器网络节点的优化空间位置配置方案,得出对运动目标运行轨迹的优化估计算法,最终得到目标运动参数估计。通过仅传送峰值功率测量的方法,降低了传感器网络内通信流量,能够很好的解决无线传感器节点通信能力有限的问题。仿真实验显示:即使在信噪比不大的情况下,用传感器网络测量峰值功率的方法进行目标运动分析已能达到相当的估计精度。更多还原

【Abstract】 Wireless location finding has emerged as an essential feature in various important applications. Wireless localization and tracking of objects of interest are two canonical issues in wireless sensor networks and roaming robots. The related digital signal processing algorithms are key factors which impact directly on positioning accuracy, stability and real-time implementation. While intensive research in this area continued to deepen, there are still many technical bottlenecks and practical problem need to solve. Especially after the UAV and wireless sensor networks appeared, such low-cost, ultra-compact, network-based detection form became widely used, the corresponding research on practical technology lags behind.This dissertation deals systematically with passive point source localization and target motion analysis algorithm in Gaussian noise environment, focusing on improving precision and robustness, finding low-cost practical solutions. The contributions can be concluded as follows:(1) Approximate maximum likelihood estimation algorithms for stationary target localization.Stationary target localization based on Angle Of Arrival (AOA) and Time-Difference Of Arrival(TDOA) is investigate systematically. Fisher Information Matrixes based on TDOA/AOA algorithms are calculated. By introducing an intermediate variable, the nonlinear equations relating estimates can be transformed into a set of equations which are linear in the unknown parameters. Under some mild conditions, the noise term behaves like normal distributed. Therefore, the solutions are presented as approximate realization of the maximum likelihood estimators(MLE). Simulations show that the proposed algorithms outperform the previous contribution.(2) Constrained WLS and wireless localization based on TDOA measurementsSource localization is studied based on assumed Gauss distributed noisy measurements of TDOA (time-difference of arrival). The solution to the constrained weighted least-squares (WLS) is derived and applied to the source localization problem based on TDOAs. Under some mild conditions, the proposed algorithm is shown to be an approximate maximum likelihood estimation (MLE) algorithm. The simulation results show that the proposed approximate MLE algorithm compares favorably with the existing solution methods.(3) Constrained Kalman Filter for Localization and Tracking Based on TDOA and DOA MeasurementsThe problem of localization and tracking of moving target is investigated based on measurements of TDOA (time-difference of arrival) and DOA (direction of arrival) for which the measurement noises are assumed to be independent and identically distributed. The problem of the constrained linear MMSE (minimum mean-squared error) estimation is formulated by employing the pseudo-measurement model that imposes a quadratic constraint on the state vector associated with the target dynamics. Randomization of the state vector for the moving target process suggests to replace the hard constraint by its expectation. We first derive a solution to a quadratically constrained MMSE estimation problem. The constrained Kalman filtering is then developed for those estimation problems involving quadratic constraints, applicable to localization and tracking of moving target based on TDOA and DOA measurements. Moreover, the constrained Kalman filter admits a simple recursive solution with complexity comparable to that of the conventional Kalman filter. A simulation example is used to illustrate our proposed constrained Kalman filter outperform EKF in localization and tracking.(4) Target Motion Analysis for smart device by using Power and Doppler measurementsTarget motion analysis for smart device is investigated based on noisy measurements of power and Doppler from a radio signal emitter. The research is motivated by the fact that Power & Doppler measurements for RF signals are much easier to obtain on small detective device. The corresponding Fisher information matrix shows that power and Doppler measurements complement each other. An iterative least-squares algorithm based on Power & Doppler measurements is developed. It is shown to be effective through a numerical simulation.(5) Target Motion Analysis Based on Peak Power Measurements Using Networked Sensors Target Motion Analysis (TMA) using a network of wireless sensors/receivers which measure the power from a mobile Radio Frequency (RF) emitter is considered. Due to the limited communication capability of each sensor node, only peak power measurements from sensor nodes are transmitted to the fusion center. We present two main results that yield the optimum sensors’ configuration such that the asymptotically achievable error variance of the target trajectory’s estimate is minimized, and we derive efficient numerical algorithms for computing the optimum estimates of the trajectory of the moving target, thus achieving the goal of TMA. Simulation experiments showed that: even in the case of small signal to noise ratio, the proposed method has been able to achieve considerable estimation accuracy.更多还原