基于状态可观测性和多模态数据PF的移动目标跟踪
Moving Target Tracking Based on State Observability and Multimodal Data PF
胡国华 1赵涓涓 2郝耀军1
作者信息
- 1. 忻州师范学院计算机系,山西忻州 034000
- 2. 太原理工大学信息与计算机学院,山西太原 030024
- 折叠
摘要
为了实现对移动目标的跟踪,提出了一种新颖的基于目标状态可观测性和多模态数据粒子滤波(Particle Filter,PF)的跟踪方案.通过部署在目标移动区域中的传感器获得跟踪目标的距离和到达方向测量值,对接收到的数据进行预处理来计算PF的观测值,以形成一个临时距离图像.通过利用状态更新函数和形成的候选图像模板确定目标状态向量;在PF器中加入额外的加权阶段,使得PF器可自适应地同步多模态数据流,以实现鲁棒的目标跟踪.仿真实验结果验证了所提方案能够有效地跟踪移动目标.
Abstract
In order to track moving targets,a novel tracking scheme based on the observability of target state and multimodal data Particle Filtering(PF)is proposed.Using the measurements of the distance and direction of arrival for the tracking target obtained by the sensors deployed in the target moving region,the particle filtering observation value is calculated by preprocessing the received data to form a temporary range image.Then the target state vector is determined by using the state update function and the formed candidate image template.In addition,an additional weighted stage is added to the particle filter,so that the particle filter can adaptively synchronize the multimodal data stream to achieve robust target tracking.The simulation results show that the proposed scheme can track the moving target effectively.
关键词
无线传感器网络/移动目标跟踪/状态向量/粒子滤波/多模态数据/传播延迟Key words
wireless sensor network/moving target tracking/state vector/PF/multimodal data/propagation delay引用本文复制引用
基金项目
山西省自然科学基金(20210302124330)
山西省高等学校科技创新项目(2019L0847)
教育部人文社会科学研究青年基金(20YJC630034)
忻州师范学院五台山文化生态研究院专项(2020133101)
出版年
2024
国家科技期刊平台
NSTL
万方数据