A target tracker using spatially distributed infrared measurements

An extended Kalman filter algorithm is designed to track a point source target in an open-loop tracking problem, using outputs from a forward-looking infrared (FLIR) sensor as measurements. The filter separately estimates the translational position changes of the target in the FLIR field of view due to two effects: actual target motion and apparent motion caused by atmospheric turbulence. A Monte Carlo analysis is conducted to determine the performance of the filter as a function of signal-to-noise ratio, target spot size, the ratio of rms target motion to rms atmospheric jitter, target correlation times, and mismatches between the true target spot size and the size assumed by the filter. The performance of the extended Kalman filter is compared to the performance of an existing correlation tracker under identical conditions. A one sigma tracking error of 0.2 and 0.8 picture elements is obtained with signal-to-noise ratios of 20:1 and 1:1, respectively. No degradation in performance is observed when the spot size is decreased or when the target correlation time is increased over a limited range, when filter parameters are adjusted to reflect this knowledge. Sensitivity analysis shows that the filter is robust to minor changes in target intensity spot size.     

改进扩展卡尔曼滤波算法的目标跟踪算法

为了获得更加理想的运动目标跟踪效果,提出了一种基于改进扩展卡尔曼滤波的目标跟踪算法。构建时间差和信号到达方向的观测方程,利用几何和代数关系化简得到伪线性模型,通过改进卡尔曼滤波算法对目标运动轨迹进行跟踪,采用仿真实验对算法性能进行测试。结果表明,相对于传统扩展卡尔曼滤波算法,在相同条件下,该算法不仅提高了目标跟踪精度,而且使目标跟踪结果更加稳定。     

Robust visual tracking control system of a mobile robot based on a dual-Jacobian visual interaction model

This paper presents a novel design of a robust visual tracking control system, which consists of a visual tracking controller and a visual state estimator. This system facilitates human–robot interaction of a unicycle-modeled mobile robot equipped with a tilt camera. Based on a novel dual-Jacobian visual interaction model, a robust visual tracking controller is proposed to track a dynamic moving target. The proposed controller not only possesses some degree of robustness against the system model uncertainties, but also tracks the target without its 3D velocity information. The visual state estimator aims to estimate the optimal system state and target image velocity, which is used by the visual tracking controller. To achieve this, a self-tuning Kalman filter is proposed to estimate interesting parameters and to overcome the temporary occlusion problem. Furthermore, because the proposed method is fully working in the image space, the computational complexity and the sensor/camera modeling errors can be reduced. Experimental results validate the effectiveness of the proposed method, in terms of tracking performance, system convergence, and robustness.     

基于改进布谷鸟粒子滤波算法的WSN目标跟踪

为了解决粒子滤波(PF)的无线传感器目标跟踪中样本贫化导致的精度较低的问题,提出了改进布谷鸟粒子滤波的WSN目标跟踪方法。通过改进布谷鸟算法的滤波算法取代粒子滤波重采样过程,主要通过改进布谷鸟算法中的搜索步长值 和发现外来鸟卵的物种的概率 的自适应调节,同时在步长更新方程中实时引入函数值的变化趋势,引导粒子整体上向较高的随机区域移动, 有效调整全局探索和局部探索适应能力、改善粒子贫化和局部极值问题,增加粒子群多样化从而提高跟踪性能。实验结果表明,改进布谷鸟粒子滤波算法重采样方法可以防止粒子的退化,增加粒子的多样性,减少跟踪误差,可以减少算法的运行时间,实时追踪性能大幅提高。与CS-PF算法和PF算法相比较,ICS-PF 算法的计算时间是最短的,ICS-PF算法的位置和速度的平均平方根误差最小(位置0.0306、0.0213、速度0.0253、0.0102),PF算法的跟踪精度是最低的,而ICS-PF跟踪精度较高,ICS-PF算法被证明具有良好的跟踪性能。     

Chance-constrained target tracking using sensors with bounded fan-shaped sensing regions

We present a robust target tracking algorithm for a mobile robot. It is assumed that a mobile robot carries a sensor with a fan-shaped field of view and finite sensing range. The goal of the proposed tracking algorithm is to minimize the probability of losing a target. If the distribution of the next position of a moving target is available as a Gaussian distribution from a motion prediction algorithm, the proposed algorithm can guarantee the tracking success probability. In addition, the proposed method minimizes the moving distance of the mobile robot based on the chosen bound on the tracking success probability. While the considered problem is a non-convex optimization problem, we derive a closed-form solution when the heading is fixed and develop a real-time algorithm for solving the considered target tracking problem. We also present a robust target tracking algorithm for aerial robots in 3D. The performance of the proposed method is evaluated extensively in simulation. The proposed algorithm has been successful applied in field experiments using Pioneer mobile robot with a Microsoft Kinect sensor for following a pedestrian.     

基于卷积特征的核相关自适应目标跟踪

目的 针对现实场景中跟踪目标的快速运动、旋转、尺度变化、遮挡等问题,提出了基于卷积特征的核相关自适应目标跟踪的方法。方法 利用卷积神经网络提取高、低层卷积特征并结合本文提出的核相关滤波算法计算并获得高底两层卷积特征响应图。采用Coarse-to-Fine方法对目标位置进行估计,在学习得到1维尺度核相关滤波器估计尺度的基础上实时更新高低两层核相关滤波器参数,以实现自适应的目标跟踪。结果 实验选取公开数据集中的典型视频序列进行跟踪,测试了算法在目标尺度发生变化、遮挡、旋转等复杂场景下的跟踪性能并与多种优秀的跟踪算法在平均中心误差、平均重叠率等指标上进行了定量比较,在Singer1、Car4、Jogging、Girl、Football以及MotorRolling视频图像序列上的中心误差分别为8.71、6.83、3.96、3.91、4.83、9.23,跟踪重叠率分别为0.969、1.00、0.967、0.994、0.967、0.512。实验结果表明,本文算法与原始核相关滤波算法相比,平均中心位置误差降低20%,平均重叠率提高12%。结论 采用卷积神经网络提取高低两层卷积特征,高层卷积特征用于判别目标和背景,低层卷积特征用于预测目标位置并通过Coarse-to-Fine方法对目标位置进行精确的定位,较好地解决了由于目标的旋转和尺度变化带来的跟踪误差大的问题,提高了跟踪性能并能够实时更新学习。在目标尺度发生变化、遮挡、光照条件改变、目标快速运动等复杂场景下仍表现出较强的鲁棒性和适应性。     

随机系统中能容忍连续丢包和测量时延的卡尔曼滤波

通过转换原线性系统到能容忍连续丢包和测量时延的随机参数系统,推导了一个最优线性滤波器.给出一个仿真例子,比较已存在的结果,仿真结果表明被提出的线性滤波器有优越的性能.然而,该滤波器不能应用于非线性系统.从应用角度,为非线性系统提出了一个增强型的滤波器.而且,该增强型的滤波器能成功地应用于不可靠的无线传感器网络场景来跟踪移动目标.这些滤波器只依靠测量值的达到概率,而不需要知道某一时刻测量是否接收.仿真说明了被提出的增强型滤波器不仅能改善实时目标跟踪的鲁棒性,而且比标准的扩展卡尔曼滤波器能够提供更精确的估计.     

基于固定点迭代的Huber鲁棒容积卡尔曼滤波算法

对于非线性系统而言,容积卡尔曼滤波（Cubature Kalman Filter,CKF）算法是处理状态估计问题的一种有效方法,并且其在高斯噪声下可以获得良好的估计性能。然而,当噪声被重尾噪声污染时,其性能通常会急剧下降。为解决此问题,将Huber方法应用于CKF框架中,取代了传统的最小均方误差（Minimum Mean Square Error,MMSE）准则,以提高算法的鲁棒性。在所提算法中,通过将量测方程线性化构造了线性回归模型,并采用固定点迭代的方法求解基于Huber方法的最小化问题。因此,推导了基于固定点迭代的Huber鲁棒CKF（FP-IHCKF）算法,在该算法中先验信息和量测信息通过Huber方法进行了重构。通过对再入目标跟踪问题进行仿真,验证了所提算法的有效性和鲁棒性。     

Interacting multiple model and sensor selection algorithms for manoeuvring target tracking in wireless sensor networks with multiplicative noise

This article addresses the problem of tracking a manoeuvring target in a wireless sensor network (WSN) consisting of distance-measuring sensor nodes. In order to cope with target manoeuvres, an interacting multiple model (IMM) filter is applied to estimate the position and velocity of the target. The distance-dependent measurement error of sensors is formulated as both additive and multiplicative noise in the observation equation. To deal with nonlinearities in the process and observation equations and also to solve the problem of multiplicative measurement noise, a new particle filter (PF)-based IMM approach is developed. Furthermore, the multiple-model posterior Cramér-Rao lower bound (PCRLB) is derived in the presence of both additive and multiplicative noise and it is used to perform a sensor selection algorithm to reduce energy consumption in WSN nodes. Simulation results show the effectiveness of the proposed IMMPF and sensor selection algorithms in target tracking.     

Interacting multiple model filter-based distributed target tracking algorithm in underwater wireless sensor networks

This paper deals with the problem of accurately tracking a single target, which has various trajectories, moving through the environment of underwater wireless sensor networks (UWSNs). This paper addresses the issues of estimating the states of the target, improving energy efficiency by using a distributed architecture. Each underwater wireless sensor node composing the UWSNs is battery-powered, so the energy conservation problem is a critical issue. This paper provides algorithms increasing the energy efficiency of each sensor node by using the proposed Wake-Up/Sleep (WUS) scheme. An interacting multiple model (IMM) filter is applied to the proposed distributed architecture in order to cope with a target maneuver. Simulation results illustrate the performance of the proposed tracking filter according to the various target maneuver patterns.     

Iterated square root unscented Kalman filter for maneuvering target tracking using TDOA measurements

This paper presents a study involving the prediction of a complicated maneuvering target, with the aim of improving the tracking performance of time difference of arrival (TDOA) tracking system for passive radar. Because of the large error caused by the complicated maneuvers and a high realtime requirement, the TDOA tracking system will take a heavy computational load. In this study, we calculate the initial position of a complicated maneuvering target using the total least square method to decrease the initial tracking error. Based on the current statistical model and the square root unscented Kalman filter, an iterated square root unscented Kalman filter (ISRUKF) is presented and an iterated termination criteria is used to reduce the linearity error for the whole iterated process. Finally, comparative simulation results are provided to demonstrate the effectiveness and applicability of the proposed method.     

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.     

粒子滤波在无线传感器网络目标跟踪中的应用

将粒子滤波（PF）算法应用于无线传感器网络（WSNS）的目标跟踪,并给出了粒子滤波实现的具体步骤。动态组织传感器网络节点成簇,实现了对网络中做匀速直线运动的单个目标的跟踪。分别采用扩展卡尔曼滤波（EKF）、无迹卡尔曼滤波（UKF）和PF算法进行了仿真试验。结果表明,在无线传感器网络目标跟踪领域,PF算法比EKF算法、UKF算法的滤波精度更高,性能更好,并且在实际应用中,由于该算法能够有效解决非线性、非高斯环境中的目标跟踪问题,实现简单而增强了可用性。     

基于多观测模型的粒子滤波头部跟踪算法

粒子滤波在非线性和非高斯问题上具有独特的优越性,但在视频跟踪过程中,其跟踪性能却在很大程度上依赖于观测模型的选择。为了解决被跟踪目标特征状态随时间变化而与粒子观测模型不匹配的问题,提出了一种新的粒子滤波算法,即将被跟踪目标的不同特征状态与粒子观测模型相结合,形成一组具有不同观测模型的粒子,并且在跟踪过程中,对应不同观测模型的粒子根据被跟踪目标所表现的特征线索的变化而相互转换,从而动态刻画了被跟踪目标特征变化的过程。实验结果表明,本算法能够有效处理由于头部旋转而导致跟踪性能下降甚至丢失跟踪目标的问题,提高了跟踪的准确性,并且具有较好的鲁棒性。     

快速尺度估计的增强型多核相关滤波算法

针对核相关滤波（KCF）跟踪算法在复杂环境下其定位性能和稳定性差的问题,提出了一种快速尺度估计的增强型多核相关滤波跟踪算法。该算法针对核相关滤波算法无法适应跟踪过程中目标尺度变化,将快速判别式尺度估计移植至核相关滤波跟踪框架,解决了跟踪过程的目标尺度问题。对于单个特征的单核相关滤波器在复杂环境中跟踪适应性差的问题,提出了一种多特征互补的多核相关滤波器。该滤波器利用KCF多通道特性以及不同特征可以描述不同信息,采用多个相同内核的线性组合,每个内核对应一个特征,并结合快速尺度估计,在保证算法实时性的同时进一步提高跟踪性能。通过在OTB2013目标跟踪数据集上进行实验,该算法与近年来性能优异的算法进行对比,结果表明,与传统的使用HOG特征的KCF算法相比精度上提高了10.9%,成功率提高了16.2%;与使用CN特征的CN2算法相比,精度上提高了20.6%,成功率提高了19.6%。实验结果表明,所提算法在目标尺度变化以及复杂环境下的跟踪效果均优于其余相关滤波算法,证明了该算法的有效性以及鲁棒性。     

Square-root unscented Kalman filtering-based localization and tracking in the Internet of Things

The Internet of Things (IoT), which is usually established over architectures of wireless sensor networks, provides an actual platform for various applications of personal and ubiquitous computing. Recently, moving target localization and tracking in an IoT environment have been paid more and more attention. This paper proposes a square-root unscented Kalman filtering (SR-UKF)-based algorithm to discover real-time location of a moving target in an IoT environment where there exist quantities of sensors. The data generated from wireless sensor nodes of the IoT make contributions to localization and tracking of the moving target. First, a least-square (LS) criterion-based mathematical model is proposed for localization initialization in an IoT scenario. Next, we employ an SR-UKF idea for the further localization and tracking. By using the data coming from sensor nodes near the target, real-time location of the moving target can be estimated by implementation of SR-UKF in an iterative fashion so as to achieve target status tracking. Simulation results show that the proposed algorithm achieves good performance in estimation of both position and velocity of the target with either uniform linear motion or variable-speed curve motion. Compared with some existing conventional extended Kalman filtering (EKF) or UKF-based methods, the proposed algorithm shows lower location/velocity estimation error under the same computational complexity, which demonstrates its potential significance in ubiquitous computing applications for an IoT environment.     

粒子滤波实现无线传感器网络目标跟踪预测

为减少无线传感器网络(WSN)目标跟踪预测误差,提出一种粒子滤波实现WSN目标跟踪预测方法;该方法采用粒子滤波获得目标运动状态,联合当前时刻目标的本地估计位置、预测速度和加速度获得下一时刻目标预测位置,预测位置可作为当前头节点唤醒所述下一时刻传感器节点的依据;结果表明,上述粒子滤波预测方法预测准确度相比线性预测方法明显提高,均方根误差RMSE减少49%;相比基于二次多项式运动建模的WSN目标跟踪预测方法,均方根误差RMSE减少6%。     

基于自适应蝙蝠粒子滤波算法的WSN目标跟踪

为了解决粒子滤波(PF)的无线传感器目标跟踪中样本贫化导致的精度较低的问题,提出了自适应蝙蝠粒子滤波的WSN目标跟踪方法。通过自适应的蝙蝠算法的滤波算法优化粒子重采样过程,结合最新的观测值定义粒子的适应度函数,引导粒子整体上向较高的随机区域移动。同时利用动态自适应惯性权重探索新的粒子位置更新为设计机制,引入动态适应惯性权重值, 有效调整全局探索和局部探索适应能力、改善粒子贫化和局部极值问题,增加粒子群多样化从而提高跟踪性能。实验结果表明,自适应蝙蝠粒子滤波算法重采样方法可以防止粒子的退化,增加粒子的多样性,减少跟踪误差,可以减少算法的运行时间,实时追踪性能大幅提高。与BA-PF算法和PF算法相比较,IBAPF 算法的计算时间是最短的,IBA-PF算法的位置和速度的平均平方根误差最小(位置0.0311、0.0202、速度0.0262、0.0101),PF算法的跟踪精度是最低的,而IBA-PF跟踪精度较高,IBA-PF算法被证明具有良好的跟踪性能。     

Event-averaged maximum likelihood estimation and mean-field theoryin multitarget tracking

This paper presents a novel type of Kalman filter for track maintenance in multitarget tracking using thresholded sensor data at high target/clutter densities and low detection levels. The filter is robust against tracking errors induced by crossing tracks, clutter, and missed detections, and the computational complexity of the filter scales well with problem size. There are two key features that differentiate this approach from earlier work. First, to reduce computational load, the filter exploits techniques from statistical field theory to simplify measurement to track association by using a mean-field approximation to sum over associations. Second, to enhance tracking of close together targets, the filter explicitly models the error correlations that occur between such target pairs. These error correlations are caused by measurement to track association ambiguities that arise when target separations are comparable to sensor measurement errors     

WSNs下一种自适应多传感器协同目标跟踪策略*

对WSNs中机动目标跟踪问题提出一种自适应多传感器协同跟踪策略.该策略能根据目标的移动位置,动态地唤醒无线传感器网络中部分传感器节点形成分簇,并选择合适的簇首和采样间隔进行目标跟踪.簇内节点通过协作感知以及测量信息融合,提高了跟踪精度,同时自适应可变采样间隔节约了通信能量和计算资源,满足了跟踪系统的实时性要求.提出了传感器网络能量均衡分配的指标,提高了网络的可靠性.由于模型的非线性和目标运动的机动性,采用IMM滤波器进行目标状态估计.仿真结果表明,与NSSS和DGSS相比,跟踪精度明显提高;与DCSS相比,在保证一定跟踪精度的同时,节约了能量消耗.