摘要
研究了非高斯噪声条件下非线性运动的目标跟踪问题。给出了非高斯噪声的两种产生方法,分别是概率分布函数法和生成数据库法,前者具有完整的数学理论支撑,后者易于操作实现。提出了一种基于置信区间的变步长重采样方法,并将其融入到粒子滤波的方法过程中。仿真表明,产生的非高斯噪声具有较好的尖峰特性,新的粒子滤波方法能够实现更高精度的目标跟踪,避免了粒子退化问题,为非线性目标跟踪问题的研究提供了一种思路。
Target tracking problem of nonlinear moving in non-Gaussian noise was studied. Two methods of generating non- Gaussian noise were given, which were probability distribution function and data base. The former was supported by strict mathematic theory, and the latter was easy to operate. Furthermore, one variational-step re-sampling method based on confidence interval was proposed, which was added into process of particle filter. Simulation shows generated non-Gaussian noise has good peak char- acteristics, new particle filter method could realize target tracking with high precision and avoid the disadvantage of particle degen- eracy, which provides new thoughts for study of nonlinear target tracking.
出处
《电子技术应用》
北大核心
2014年第8期129-132,共4页
Application of Electronic Technique
基金
陕西省电子信息综合集成重点实验室项目(201107Y16)
关键词
非高斯噪声
目标跟踪
置信区间
重采样
粒子滤波
non-Gaussian noise
target tracking
confidence interval
re-sampling
particle filter
