改进的卡尔曼滤波与均值漂移目标跟踪算法

为了实现更加理想的运动目标跟踪,提出了改进的卡尔曼滤波与均值漂移目标跟踪算法。该算法采用粒子滤波与Kalman滤波相结合实现非线性滤波,首先利用粒子滤波对运动目标的状态变量进行估计,然后对目标估计状态进行卡尔曼滤波,解决观测方程为非线性的问题。最后利用Mean-shift算法进一步聚类粒子,使抽样的粒子集更符合实际的目标概率模型,从而增加有效粒子数目,减少粒子退化。经过预测迭代,从而达到对运动目标运行轨迹的修正,并采用仿真实验进行算法性能测试。结果表明,相对于其他算法或者是传统算法,在相同的条件下该算法不仅提高了目标跟踪的精度,并且降低了计算复杂度,实时性较好。     

一种改进的容积信息粒子滤波多目标跟踪算法

为了解决传统的自举式粒子滤波采用状态转移密度作为重要性采样函数呈现粒子退化、跟踪精度差等问题,文中采用后验概率作为重要性采样函数并提出其估计方法.通过容积信息滤波和Gating技术,估计重要性采样函数对应的均值和方差,完成重要性采样函数设计.使用改进后的粒子滤波方法,在非线性场景下对多个目标的状态和目标数目进行了估计.仿真结果表明,本算法在非线性多目标跟踪场景下具有估计精度高、稳定性好的优点.     

基于粒子滤波的非线性目标跟踪算法研究

针对非线性非高斯的目标跟踪,传统的卡尔曼滤波和扩展卡尔曼滤波等算法将会出现滤波精度下降甚至发散的现象,提出了采用粒子滤波算法来解决非线性滤波问题;粒子滤波方法作为一种基于贝叶斯估计的非线性滤波算法,在处理非高斯非线性时变系统的参数估计和状态滤波问题方面有独到的优势,但是存在运算量大和实时性差的问题,因此提出了基于EKF的扩展粒子滤波;仿真结果表明：在强非线性非高斯环境下,PF算法的跟踪性能优于EKF算法,基于EKF的扩展粒子滤波能够取得较好的跟踪精度,并且能够有效的减少粒子滤波的运算量。     

基于无迹卡尔曼滤波和权值优化的改进粒子滤波算法

针对传统粒子滤波面临的重要密度函数的选取和粒子多样性丧失引起的样本贫化问题,提出基于无迹卡尔曼滤波和权值优化的改进粒子滤波算法。与传统的粒子滤波算法相比,有两点改进:首先该算法采取无迹卡尔曼滤波产生建议分布函数;其次,在重采样过程,提出基于权值优化的改进重采样算法来增加粒子的多样性。仿真结果表明,改进算法降低了粒子滤波算法的粒子退化程度并避免样本贫化现象的出现,更加接近真实值,提高了跟踪精度。     

基于混合卡尔曼滤波的磁偶极子目标跟踪

针对经典粒子滤波(PF)对磁偶极子跟踪高维模型估计时存在的"维数灾"问题,提出一种基于混合卡尔曼滤波(MKF)的方法进行求解。该方法通过导出磁偶极子跟踪的条件线性高斯模型,利用混合卡尔曼滤波算法,将状态变量分离为线性部分和非线性部分;利用卡尔曼滤波和粒子滤波进行耦合处理,克服了经典粒子滤波直接用于磁偶极子跟踪出现的精度不高和滤波发散问题。通过仿真实验将该方法与基于PF的磁偶极子目标跟踪方法进行了对比,结果表明算法精度较高,且性能稳定。     

基于辅助模型的改进粒子滤波算法

针对非线性、非高斯系统的目标跟踪精度不太高这一问题,提出一种改进Sigma粒子滤波算法(MSP-PF)。该算法是由主模型产生第一个粒子,剩余的粒子则由辅助模型和平方根无迹卡尔曼滤波(SRUKF)来递归生成,辅助模型使粒子的观测信息得到充分有效地利用,解决了粒子滤波算法所面临的粒子退化和匮乏问题。仿真表明,提出的改进Sigma粒子滤波算法(MSPPF)的估计性能要明显优于粒子滤波(PF)、无迹粒子滤波(UPF)、Sigma粒子滤波算法(SPPF)。     

基于粒子滤波的移动机器人跟踪定位算法

针对移动机器人在非线性系统中匀速运动的实时跟踪定位问题,对基于粒子滤波的移动机器人跟踪定位算法进行研究.在控制系统中引入粒子滤波算法,并在此基础上引入退化因子,最后进行仿真验证.结果表明:该算法能避免粒子滤波算法在非线性系统的失效和单样本等问题,改善移动机器人在跟踪定位系统中的非高斯噪声影响,明显提高移动机器人定位精度.     

闪烁噪声下的机动目标跟踪算法研究

标准的粒子滤波能够较好的解决闪烁噪声下机动目标跟踪问题,但在采样过程中有时会出现严重的退化现象。针对这种现象,文中提出了一种改进的粒子滤波算法,在采样过程中退化检测,若粒子退化严重,则重新采样,直到退化在允许范围之内。仿真结果表明,在高斯环境下,标准的粒子滤波和改进的粒子滤波性能相近,都优于不敏卡尔曼滤波,但在闪烁噪声环境下,其性能明显高于不敏卡尔曼滤波和标准的粒子滤波,但耗费时间长。     

基于重采样粒子滤波的目标跟踪算法研究

基于传统粒子滤波的机动目标跟踪方法针对非线性、非高斯系统有较好的估计性能,但是存在粒子退化现象。利用残差重采样算法,可以有效克服粒子滤波的退化问题。本文针对残差重采样算法作进一步研究,提出了一种改进的残差重采样粒子滤波算法。该方法在残差重采样基础上进行改进,可以避免残差重采样中关于残留粒子的重采样问题,在保证精度的前提下提高运行效率,减少运算复杂程度。仿真实验结果表明该算法与残差重采样粒子滤波相比提高了目标跟踪的实时性,并且随着粒子数的增加,这种优势表现得更加明显。     

基于平方根卡尔曼滤波的粒子滤波算法

在普通粒子滤波器中,基于先验概率的重要性密度不能容纳最新测量信息,导致跟踪精度难以提高。针对该问题,给出一种基于平方根卡尔曼滤波(SRUKF)的新型粒子滤波算法(SRUPF)。该算法以普通粒子滤波器(PF)为基础,运用SRUKF生成重要性密度。与运用先验知识生成重要性密度的普通粒子滤波器不同,SRUPF的重要性密度中包含了最新的观测信息,从而能够更好地逼迫状态变量的分布规律。此外,由于SRUPF在计算重要性密度时不需要在每一个迭代步骤都对状态协方差阵进行分解,因而SRUPF比PF具有更好的数值稳定性。在非线性测角跟踪问题中的应用表明:SRUPF滤波器的跟踪精度优于PF和SRUKF。     

Stability of the Third Generation Intensifier under Illumination

The stability under illumination of transmission-mode GaAs photocathode sealed in the third generation intensifier is investigated by use of spectral response testing instruments. The variations of spectral response with the illumination times under weak and intense illumination are compared. The variations of photoemission performance parameters are also characterized. The results show that during initial several weak illuminations photocathode behaves no evident decay and a maximum sensitivity is achieved, while under intense illumination the sensitivity of photocathode begin to decrease largely at the first illumination. The calculated performance parameters show that the variation of surface escape probability with illumination times is a direct cause of instability of photocathode. It is also found that under intense illumination peak wavelength is moved towards short-wave and peak response is decreased, which shows that the ability of long-wave response of photocathode is decreased.     

Passive Acoustic Radar for Detecting Supersonic Cruise Missile

A Passive Acoustic Radar is presented as a necessary complement to electromagnetic wave radar, which will be expected to be an effective means for detecting cruise missiles. Acoustic characteristics of supersonic flying projectiles with diverse shapes are expounded via experiment. It is pointed out that simulation experiment could be implemented using bullet or shell instead of cruise missile. Based on theoretical analysis and experiment, the ““acoustic fingerprint““ character of cruise missile is illustrated to identify it in a strong noise environment. After establishing a locating mathematical model, the technique of acoustic embattling is utilized to resolve a problem of confirming the time of early-warning, considering the fact that velocity of sound is much slower than that of light. Thereby, a whole system of passive acoustic radar for detecting supersonic cruise missile is formed.     

Design and Analysis of Test Schemes of Static Postponing Time on Shrapnels Used to Riot Control

The test scheme of static postponing time given in our actual national military test standard on shrapnels used to control riot is a nine-point test scheme on the combined action of three kinds of temperatures and three kinds of pressures,the consumed ammunitions are more excessive. Statistic analysis and tentative checkout about a lot of test data are done,feasibility gists are put forward for optimizing of the test design scheme. The optimizing design and data analysis of test scheme of the item are done by managing uniformity design theory, two scientific and reasonable six-point test schemes are confirmed. The feasibility and reliability of the optimizing design schemes put forward above are proved ulteriorly by test validating. The gained schemes not only have good design uniformity and little ammunition wastage and meet the test demand, but also have better forecast ability for the result data of other points using the mathematic models from the actual test points.     

Super-resolution Restoration of Remote-sensing Images

A novel image restoration scheme, which is super-resolution image restoration algorithm Poisson-maximum-afterword-probability based on Markvo constraint （MPMAP） combined with evaluating image detail parameter D, has been proposed. The advantage of super-resolution algorithm MPMAP incorporated with parameter D lies in the fact that superresolution algorithm MPMAP model is discrete, which is in accordance with remote-sensing imaging model, and the algorithm MPMAP is proved applicable to linear and non-linear imaging models with a unique solution when noise is not severe. According to simulation experiments for practical images, super-resolution algorithm MPMAP can retain image details better than most of traditional restoration methods; at the same time, the proposed parameter D can help to identify real point spread function （PSF） value of degradation process. Processing result of practical remote-sensing images by MPMAP combined with parameter D are given, it illustrates that MPMAP restoration scheme combined PSF estimation has a better restoration result than that of Photoshop processing, based on the same original images. It is proved that the proposed scheme is helpful to offset the lack of resolution of the original remote-sensing images and has its extensive application foreground.     

Design and Simulation of TF and TA Controller for Missile

A nonlinear terrain following(TF) and terrain avoidance(TA) controller is proposed for missile control systems. Based on classical TF algorithm (adaptive angle method), a new method for TF controller is proposed by using angle of attack. A method of obtaining terrain outline data from digital elevation map (DEM) for TF control is discussed in order to save store space. A TA algorithm is proposed by using bank-to-turn technique. The block control model, which is suitable for backstepping design, is given for nonlinear model of missile. Making full use of the characteristics of the system and combining block control principle and backstepping technique, a robust controller design method is proposed. Uncertainties in every sub-block are allowed, and can be canceled by using the idea of nonlinear damping. It is proved that the state tracking errors are converged to a neighborhood of the origin exponentially. Finally, nonlinear six-degree-of-freedom simulation results for the missile model are presented to demonstrate the effectiveness of the proposed control law.     

Design of Non-fragile Satisfactory Estimator for Linear Continuous Perturbed Stochastic Systems

The design problem of non-fragile estimator is addressed for a class of perturbed linear continuous systems. The perturbations occur on the plant and estimator parameters. The estimator designed should force the error system to achieve the desired decay rate and force the steady error variance less than the specified upper bound irrdevancy of the admissible plant perturbations and estimator perturbations. Consistency problem of the decay rate with the variance upper bound is first considered via linear matrix inequality （LMI） approach. The solution of the estimator parameter under specifications to be consistent is then discussed. The consistency condition of specifications and estimator parameter solution are transformed to feasible or minimum problems subject to a set of LMI respectively. The method is illustrated by a numerical example.     

An Improved Scene-based Nonuniformity Correction Algorithm for Infrared Focal Plane Arrays Using Neural Networks

The improved scene-based adaptive nonuniformity correction （NUC） algorithms using a neural network （NNT） approach for infrared image sequences are presented and analyzed. The retlna-like neural networks using steepest descent model was the first proposed infrared focal plane arrays （IRFPA） nonuniformity compensation method, which can perform parameter estimation of the sensors over time on a frame by frame basis. To increase the strength and the robustness of the NNT algorithm and to avoid the presence of ghosting artifacts, some optimization techniques, including momentum term, regularization factor and adaptive learning rate, were executed in the parameter learning process. In this paper, the local median filtering result of Xij （ n ） is proposed as an alternative value of desired network output of neuron Xij （ n ）, denoted as Tij （ n ）, which is the local spatial average of Xij （ n ） in traditional NNT methods. Noticeably, the NUC algorithm is inter-frame adaptive in nature and does not rely on any statistical assumptions on the scene data in the image sequence. Applications of this algorithm to the simulated video sequences and real infrared data taken with PV320 show that the correction results of image sequence are better than that of using original NNT approach, especially for the short-time image sequences （several hundred frames） subjected to the dense impulse noises with a number of dead or saturated pixels.     

Piecewise Filter of Infrared Image Based on Moment Theory

The disadvantages of IR images mostly include high noise, blurry edge and so on. The characteristics make the existent smoothing methods ineffective in preserving edge, To solve this problem, a piecewise moment filter （PMF） is put forward. By using moment and piecewise linear theory, the filter can preserve edge. Based on the statistical model of random noise, a related-coefficient method is presented to estimate the variance of noise. The edge region and model are then detected by the estimated variance. The expectation of first-order derivatives is used in getting the reliable offset of edge. At last, a fast moment filter of double-stair edge model is used to gain the piecewise smoothing results and reduce the calculation. The experimental result shows that the new method has a better capability than other methods in suppressing noise and preserving edge.     

Bistatic Sonar Localization Based on Best Linear Unbiased Estimation

A best linear unbiased estimation （BLUE） algorithm for bistatic sonar localization is proposed. The Cramer-Rao bound for bistatic Sonar and the geometrical dilution of precision （GDOP） in different conditions are given. The simulation results show that the location accuracy of BLUE algorithm is higher than the weighted least square method.     

Research on Gear-broken Fault Diagnosis in a Tank Gearbox

A fault diagnosis method of working position gear in a tank gearbox is put forward based on simulating the fault of working position gear in an actual tank, extracting the envelope of vibration signal by Hilbert transformation amplitude demodulation method, and zooming the low-frequency band to envelope signal.