多源传感器数据融合自主稳定跟踪算法研究

发布时间：2018-03-01 08:01

本文关键词： 多源传感器数据融合自主稳定跟踪抗干扰　出处：《中国科学院长春光学精密机械与物理研究所》2017年硕士论文　论文类型：学位论文

【摘要】：光学测量系统是靶场试验中用于目标捕获、跟踪和测量的常用设备。主要通过光学成像技术对空间飞行目标进行跟踪测量,获得靶场目标的高精度弹道数据和飞行实况影像,为武器鉴定和故障分析提供重要的数据。集多种工作波段光学传感器为一体的光电经纬仪跟踪系统,充分利用多传感器信息进行目标跟踪,可以达到单一传感器跟踪方式无法企及的效果。目前,多传感器数据融合技术大多用于提高目标定位精度,研究该技术在提高跟踪稳定性和抗干扰能力方向上应用的文献较少。本文以靶场光电跟踪系统为研究平台,致力于通过多源传感器的数据融合技术提高跟踪系统的抗干扰能力和跟踪稳定性。本文主要研究内容和成果如下。1.对多传感器数据融合自主稳定跟踪的相关理论进行了阐述,包括多波段和多参数光学传感器数据的融合原理、常用的多传感器数据融合目标跟踪算法以及光电自主稳定跟踪的意义和目的等。根据实际跟踪场景中各光学传感器的图像特征和跟踪数据特点,总结出了光电经纬仪自主稳定跟踪的常见影响因素和对应的数据变化特点。2.研究并改进了具有容错性的多源传感器数据融合算法。介绍了光学传感器脱靶量预处理和经纬仪过零点处理的解决方案。改进了原基于估计误差协方差的数据融算法,提出最小二乘曲线拟合与记忆衰减因子相结合的误差估计方法,提高了误差估计的准确性和实时性。并在融合算法基础上加入了去野值和防发散的处理过程。3.研究并提出了基于数据融合的多传感器自主稳定跟踪算法。根据跟踪系统的主要目的,将自主跟踪策略分为目标优先多传感器自主选择跟踪算法和精度优先多传感器自主协同跟踪算法。4.对数据融合算法和目标优先自主选择跟踪算法的实验结果进行了仿真分析和对比。仿真结果表明,改进后的数据融合算法在容错性能和跟踪精度上均有一定提高,自主选择跟踪算法可达到满意的数据选择效果。结合实际跟踪数据,分析了数据融合算法和两种自主稳定跟踪算法的实验结果。本文全面完整地论述了多传感器数据融合自主稳定跟踪算法的目的、原理和实现方法,最后将本文的数据融合及自主跟踪算法应用在某靶场大型光学测量设备中,证明了上述算法的有效性和实用性。
[Abstract]:Optical measurement system is commonly used in target acquisition, tracking and measurement in range test. To provide important data for weapon identification and fault analysis. The photoelectric theodolite tracking system, which integrates optical sensors in various working bands, makes full use of multi-sensor information for target tracking. At present, the multi-sensor data fusion technology is mostly used to improve the accuracy of target location. The application of this technique in improving tracking stability and anti-jamming ability is less. In this paper, the photoelectric tracking system of shooting range is used as the research platform. The main contents and achievements of this paper are as follows: 1. The related theory of autonomous stable tracking of multi-sensor data fusion is described. Including multi-band and multi-parameter optical sensor data fusion principle, The commonly used multi-sensor data fusion target tracking algorithm and the significance and purpose of optoelectronic autonomous and stable tracking. According to the image features and tracking data characteristics of each optical sensor in the actual tracking scene, This paper summarizes the common influencing factors and the corresponding data change characteristics of the photoelectric theodolite autonomous stable tracking. 2. The data fusion algorithm of multi-source sensor with fault tolerance is studied and improved. The optical sensor miss distance preposition is introduced. An improved data fusion algorithm based on estimation error covariance. An error estimation method combining least square curve fitting with memory attenuation factor is proposed. The accuracy and real-time performance of error estimation are improved, and the processing process of outliers and anti-divergence is added to the fusion algorithm. 3. A multi-sensor autonomous stable tracking algorithm based on data fusion is studied and proposed. The main purpose of the tracer system, The autonomous tracking strategy is divided into target first multi-sensor autonomous selection tracking algorithm and precision priority multi-sensor autonomous cooperative tracking algorithm .4.The experimental results of data fusion algorithm and target first autonomous selection tracking algorithm are carried out. The simulation results show that, The improved data fusion algorithm has improved both fault-tolerant performance and tracking accuracy, and the self-selection tracking algorithm can achieve satisfactory data selection effect. The experimental results of the data fusion algorithm and two autonomous stable tracking algorithms are analyzed. The purpose, principle and implementation of the multi-sensor data fusion autonomous stability tracking algorithm are discussed in this paper. Finally, the data fusion and autonomous tracking algorithms are applied to a large optical measuring equipment in a range, which proves the validity and practicability of the above algorithms.
【学位授予单位】：中国科学院长春光学精密机械与物理研究所
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41;TP212.9

【参考文献】