New results in optimizing surveillance system tracking and data correlation performance in dense multitarget environments

Tracking performance in surveillance systems depends on two interrelated functions: track updating, the process of incorporating a new measurement into the track to update the system state estimate, and return-to-track correlation, the process of selecting which sensor return, if any, to use for track updating. Because of the presence of a number of targets in the same vicinity and the existence of clutter and false alarms, the correlation function is generally performed imperfectly. Since typical tracking filters such as the Kalman filter do not account for such correlation errors, degraded performance often results as well as unreliable and optimistic estimates of tracking accuracies. This paper examines and provides for optimizing the overall tracking process considering both the correlation and track update functions and their interaction. General equations for tracking performance of any arbitrary tracking filter used with a broad class of correlation algorithms in dense multitarget environments are developed. A new reoptimized tracking filter is derived which provides, from among a general class of tracking filters using a priori information on sensor return statistics, optimal performance in such environments and which reduces to the Kalman filter when environmental effects are eliminated. The new filter is compared parametrically to both the standard Kalman filter and a computationally simpler version of the optimal filter in terms of tracking accuracy and reliability of the calculated covariance matrix, over a spectrum of environmental conditions. At high densities of sensor returns, the new filter provides considerably improved tracking performance as well as uniquely reliable estimates of this performance.     

自适应门限GM-CPHD多目标跟踪算法

带有势估计的高斯混合概率假设密度滤波（GM-CPHD）作为一种杂 波环境下目标数可变的检测前跟踪方法,将复杂的多目标状态空间的运算转换为单目标状态 空间内的运算,有效避免了多目标跟踪中复杂的数据关联问题,但该方法的计算复杂度与观 测数的3次方成正比,在密集杂波情况下计算量十分巨大。针对该方法计算复杂度高的问题 ,提出利用一种最大似然自适应门限的快速算法,该算法首先利用自适应门限对观测进 行处理,然后仅利用处于门限内的有效观测进行GM-CPHD算法的更新步计算,大大降低了算 法的计算复杂度。实验结果证明,本文方法在有效降低计算复杂度的同时,在多目标跟踪效 果 方面与GM CPHD相当,优于GM-PHD滤波算法。     

A Bayesian solution and its approximations to out-of-sequence measurement problems

Target tracking using delayed, out-of-sequence measurements is a problem of growing importance due to an increased reliance on networked sensors interconnected via complex communication network architectures. In such systems, it is often the case that measurements are received out-of-time-order at the fusion center. This paper presents a Bayesian solution to this problem and provides approximate, implementable algorithms for both cluttered and non-cluttered scenarios involving single and multiple time-delayed measurements. Such an approach leads to a solution involving the joint probability density of current and past target states.In contrast, existing solutions in the literature modify the sensor measurement equation to account for the time delay and explicitly deal with the resulting correlations that arise in the process noise and current target state. In the Bayesian solution proposed in this paper, such cross correlations are treated implicitly. Under linear Gaussian assumptions, the Bayesian solution reduces to an augmented state Kalman filter (AS-KF) for scenarios devoid of clutter and an augmented state probabilistic data association filter (AS-PDA) for scenarios involving clutter. Computationally efficient versions of AS-KF and AS-PDA are considered in this paper. Simulations are presented to evaluate the performance of these solutions.     

基于时间错序量测的多目标多传感器跟踪算法研究

在多传感器多目标跟踪系统中,经常有来自同一目标的量测到达融合中心时存在时间先后顺序上的混乱,被称为时间错序量测(Oosm);通常,现有的跟踪算法都是假设理想目标的观测值不混乱;现实中,可能错过的目标探测随意混乱,因而,滤波器不得不处理起因未知的量测,那么针对顺序量测的传统滤波器,例如KF,在此就不能直接使用;通过基于一些特殊矩阵非单一假设的经济存储和能效估计介绍了全局最优Oosm刷新算法,并结合概率数据关联PDA到Oosm刷新算法中;仿真结果显示Oosm刷新的PDA滤波器在性能上优于忽略Oosm的PDA滤波器,还就关于杂波中多目标跟踪如何通过JPDA结合oosm刷新算法展开讨论.     

Research on tracking of maneuvering multi-target based on bionics for IRST system

This paper mainly studies bearing-only target tracking based on bionics for IRST system. Some solutions for the key problem are presented in order to apply in an actual bearing-only engineering system. They include sensor technology, measurement pretreatment technology, association gate technology, data association technology, state filtering technology, etc. The premise of these new approaches is designing an effective sensor system which can reliably search and track targets in a large range. Then, it is important to improve the confirming efficiency of the real target and limit false track overextension with the dense clutter. Then, tracking processing needs a precise target initialization information and association information between the existing target and isolated measurement. At the same time, the threat level of the bearing-only target needs to be estimated based on limited bearing-only information. Finally, aiming at unrecognized model and complex maneuvering motion for bearing-only target in polar coordinates, an effective approach of state filtering algorithm with appropriate computation cost will be given. The application of the proposed approach in an actual engineering system proves its effectiveness and practicability.     

一种面向互联概率加权的JPDA多传感器数据融合方法

针对单传感器联合概率数据互联(Joint Probabilistic Data Association, JPDA)在复杂环境下难以跟踪多个目标的问题,提出一种基于JPDA量测目标互联概率统计加权并行式和序贯式多传感器数据融合方法。首先,给出单传感器JPDA算法。然后,介绍多传感器JPDA数学模型,基于这一模型,使用互联概率加权,推导并行式和序贯式多传感器数据融合公式,这对多传感器数据融合有一定指导意义。最后,对单传感器JPDA方法在不同杂波密度、不同过程和不同观测噪声下目标跟踪的距离RMSE进行仿真,结果表明,随着这3项指标皆增大,目标距离RMSE增大;同时,对本文的2类多传感器JPDA方法与其他几类跟踪方法在数据集PETS2009下有关行人跟踪性能进行仿真,结果表明,本文并行式和序贯式多传感器JPDA方法相较于其他方法在跟踪准确性、跟踪位置准确性、航迹维持以及航迹遗失上皆为最优,而且序贯式融合略优于并行式多传感器JPDA。     

采用序贯滤波的红外/ 雷达机动目标跟踪

当机动目标状态是非平稳和非线性时,红外传感器和雷达的目标状态方程和量测方程都是非线性和非高斯的,为了解决经典跟踪算法的残差较大或发散的问题,提出一种新的融合跟踪算法:在对红外传感器和雷达的量测数据进行时间对准和同步融合后,将融合后的量测数据送入重抽样粒子滤波器进行处理以预测和跟踪机动目标.最后给出了一个仿真跟踪实例,并与同类多雷达跟踪的效果进行了比较,说明了异类融合跟踪优于同类多雷达融合跟踪.     

基于改进PDA的弱点状多目标交叉运动跟踪算法

采用独立跟踪区域的划分和公共量测点数据的去藕聚类技术,将原本只适用于单目标跟踪的概率数据关联（PDA）算法改造成能够在强杂波环境中跟踪多个点状目标交叉运动的情况。该算法比传统基于JPDA（联合数据关联）的多目标跟踪算法的计算量和复杂度都小。仿真试验表明,该跟踪算法具有高精度的跟踪性能。     

近邻传播观测聚类的多扩展目标跟踪算法

由于传感器分辨率高或目标存在多个反射源等原因,一个目标可以同时产生多个观测数据,对于解决这种扩展目标的跟踪问题,概率假设密度(PHD)滤波算法是一种有效的方法.针对扩展目标概率假设密度滤波算法中观测集合划分,提出一种利用近邻传播聚类方法进行观测集合划分的多扩展目标跟踪算法.实验结果表明,所提出的方法不但能够获得正确的划分观测集合,而且计算复杂度较已有划分方法有较大降低,同时在多目标跟踪效果方面优于已有算法.     

The optimal order of processing sensor information in sequentialmultisensor fusion algorithms

We examine the order of sensor processing in the sequential multisensor probabilistic data association (MSPDA) filter for target tracking applications. If two sensors of different qualities are used in the sequential MSPDA filter, the r.m.s. position error is generally smaller when the worse sensor is processed first. This finding regarding the order of sensor processing is supported by simulations of a target tracking system, and by analysis of firstthrough sixth-order target process models using the modified Riccati equation     

Tracking in a cluttered environment with probabilistic data association

This paper presents a new approach to the problem of tracking when the source of the measurement data is uncertain. It is assumed that one object of interest (‘target’) is in track and a number of undesired returns are detected and resolved at a certain time in the neighbourhood of the predicted location of the target's return. A suboptimal estimation procedure that takes into account all the measurements that might have originated from the object in track but does not have growing memory and computational requirements is presented. The probability of each return (lying in a certain neighborhood of the predicted return, called ‘validation region’) being correct is obtained—this is called ‘probabilistic data association’ (PDA). The undesired returns are assumed uniformly and independently distributed. The estimation is done by using the PDA method with an appropriately modified tracking filter, called PDAF. Since the computational requirements of the PDAF are only slightly higher than those of the standard filter, the method can be useful for real-time systems. Simulation results obtained for tracking an object in a cluttered environment show the PDAF to give significantly better results than the standard filter currently in use for this type of problem.     

基于MMPHDF的多机动目标联合检测、跟踪与分类算法

针对场景中存在新目标出现、旧目标消失(即目标数目变化)和密集杂波的复杂情形,利用多模型概率假设密度滤波器(MMPHDF)在多机动目标联合检测与跟踪上的优势,加入类别辅助信息,提出了一种多机动目标联合检测、跟踪与分类算法.该算法的基本思想是在MMPHDF中用属性向量扩展单目标状态向量,用位置和属性的组合测量似然函数代替单目标位置及杂波位置测量似然函数,提高了不同类目标与杂波测量间的鉴别能力,从而改善了目标数目及状态的估计精度;在更新目标状态后,对目标属性信息进行更新,更为精确的目标数目及状态估计又保证了目标分类性能.本文给出了该算法的粒子实现方法.仿真结果验证了上述结论.     

On the divergence of information filter for multi sensors fusion

The paper deals with the divergence of information filter in multi sensor target tracking problem using bearing only measurements. Information filter has a number of advantages in terms of computational requirements over Kalman filter for target tracking applications. Compared to Kalman filter it also has the advantage that one can start estimation even without an initial estimate. But this filter is seen to diverge after tracking for a short period of time, even when the target is moving at a constant velocity. A technique to overcome this problem has been discussed in this paper. The information update equations of the conventional information filter are modified in terms of fuzzy function of error and change of error, and the results have been found to be encouraging. The efficacy of the technique in preventing divergence is demonstrated in the context of tracking a maneuvering target also.     

多传感器多目标跟踪的粒子PHD滤波算法

针对单传感器跟踪系统的缺陷,提出了基于粒子概率假设密度(PHD)滤波的多传感器多目标跟踪算法.这种算法不仅避免了多传感器多目标跟踪的数据关联问题,而且在漏检、目标密集、航迹交叉、小范围内目标数多的杂波环境下能够稳定、精确地估计目标状态和目标数.仿真实验比较了单传感器粒子PHD滤波与多传感器的粒子PHD滤波的跟踪性能,验证了该方法的跟踪性能和精度.     

Simulation-based Optimal Design of α-β-γ-δ Filter

The existing third-order tracker known as α-β-γ filter has been used for target tracking and predicting for years. The filter can track the target's position and velocity, but not the acceleration. To extend its capability, a new fourth-order target tracker called α-β-γ-δ filter is proposed. The main objective of this study was to find the optimal set of filter parameters that leads to minimum position tracking errors. The tracking errors between using the α-β-γ filter and the α-β-γ-δ filter are compared. As a result, the new filter exhibits significant improvement in position tracking accuracy over the existing third-order filter, but at the expense of computational time in search of the optimal filter. To reduce the computational time, a simulation-based optimization technique via Taguchi method is introduced.     

Fairness-based joint call admission control for heterogeneous wireless networks: an SMDP approach

In heterogeneous wireless networks, both terminal heterogeneity and network heterogeneity give rise to the fairness problem of resource allocation. Due to the capability of exploiting the resources of multiple networks, the behavior of multi-mode terminals will have a great effect on single-mode terminals, and this influence becomes more severe when considering the different demands of different traffic. In this article, we propose a novel joint call admission control （JCAC） scheme to address this problem. The JCAC problem is modeled as a semi-Markov decision process （SMDP） with the aim of maximizing the average network revenue under tile constraints of the fairness among different terminals and traffic classes. Based on the SMDP, we design an algorithm to achieve a good tradeoff between revenue and fairness by dynamically adjusting the threshold of fairness constraints imposed on heterogeneous terminals. Simulation results show that the proposed scheme can significantly improve the fairness among heterogeneous terminals and guarantee the priority and fairness among different traffic classes with little loss of network revenue compared with other schemes.     

Tracking in clutter with strongest neighbor measurements. I.Theoretical analysis

When tracking a target in clutter, a measurement may have originated from either the target, clutter, or some other source. The measurement with the strongest intensity (amplitude) in the neighborhood of the predicted target measurement is known as the “strongest neighbor” (SN) measurement. A simple and commonly used method for tracking in clutter is the so-called strongest neighbor filter (SNF), which uses the SN measurement at each time as if it were the true one. The paper deals with tracking in clutter with the SN measurements. It presents analytic results, along with useful comments, for the SN measurement and the SNF, including the a priori and a posteriori probabilities of data association events, the conditional probability density functions and the covariance matrices of the SN measurement, and various mean-square-error matrices of state prediction and state update. These results provide valuable insight into the problem of tracking in clutter and theoretical foundation for the development of improved tracking algorithms, for performance analysis, prediction, and comparison of tracking with the SN measurements, and for solving some important detection-tracking problems, such as the optimal determination of the detection threshold and gate size     

TDOA/FDOA based target tracking with imperfect position and velocity data of distributed moving sensors

This paper addresses the target tracking problem using time difference of arrival (TDOA) and frequency difference of arrival (FDOA) measured by moving sensor network whose position and velocity are noise contaminated. It is a known fact that the existing approaches to this problem still have two unsolved technical issues; the unsatisfactory convergence behavior of the tracking filter mainly caused by severe nonlinearity of the problem itself and the tracking performance degradation due to the sensor position and velocity errors. In order to resolve these matters radically, the given target tracking problem is formulated as the robust state estimation problem of the linear system with stochastic uncertainties in its measurement matrix and solved by using the robust Kalman filter theory. The proposed scheme enables us to overcome the inherent limitations of the conventional nonlinear filters for its linear filter structure. It can also prevent the performance degradation due to imperfect sensor position and velocity information. Through the simulations, the effectiveness and reliable target tracking performance of the proposed method are demonstrated.     

基于概率假设密度的近邻目标快速跟踪算法

针对杂波场景中多个目标相互近邻时,标准概率假设密度滤波器难以正确估计目标状态且计算复杂度高等问题,提出一种基于概率假设密度的近邻目标快速跟踪算法。所提算法首先采用自适应门技术从传感器观测集中划分出源于真实目标的观测集,随后利用真实目标观测集来更新预测强度,最后应用检测导向的近邻目标权重校正方法有选择地重分配各离散时刻后验强度中不精确的分量权重。实验结果表明,所提算法不仅具有高效率的目标跟踪,而且具有良好的鲁棒性。     

机载雷达辅助无源传感器对杂波环境下机动目标跟踪

机载雷达辅助无源传感器对目标协同跟踪具有重要战术作用,而当前相关算法模型较为简单。为了贴近工程实际,提出一种机载雷达辅助无源传感器对杂波环境下机动目标的跟踪算法。该算法考虑了地球曲率和载机时变姿态等因素的影响,基于地心地固(ECEF)坐标系,联合交互多模型(IMM)和概率数据关联(PDAF)方法,以综合预测协方差的迹为控制变量来管理机载雷达的开关机。仿真结果表明,通过选择合适的控制门限,在节约辐射能量、提升生存能力的同时算法的跟踪性能并无明显下降,从而表明了所提出算法的有效性。