面向高效实时目标监测的阵列化传感器网络体系关键技术研究

An innovative array-forming sensor network is proposed to obtain efficient real-time target monitoring with Wireless Sensor Networks, which is inspired by array technologies and leverages modern compressive sampling theory. Technologies for real-time assurance and strategies for node scheduling are studied to facilitate the development of a distributed wireless sensor array network under resource limitations.On strategies for real-time assurance, compressive sampling and load-adaptive network protocols will be studied to build a real-time information collecting and processing framework. The performance for real-time will also be analyzed. In study of system scheduling, different modes will be designed for different requirements in order to achieve an optimum node scheduling strategies. Especially, a tradeoff strategy will be studied for occasions when resources are severely restricted. Considering the requirements for high-precision target monitoring, new target tracking framework and algorithms suitable for our system is studied using direct positioning methods, which is based on data compression. On the basis of theoretical researches, a prototype system for array-forming sensor network system will be developed to verify our models and algorithms, as well as solving the problems such as time synchronization, self localization. This project provide theoretical framework as well as experimental verification for large-scale real-time wireless sensor network to solve the problem of restricted real-time processing ability and resource, which is significant in its way to practical application.

面向高效实时的目标监测，本课题在阵列技术思想的启发下，以最新的压缩采样技术为支持，提出了一种新型的阵列化传感器网络体系架构，重点研究资源受限的实时保障和系统调度策略。实时保障方面，利用压缩采样技术和负载自适应的网络通信协议，构建网络化信息实时采集传输及处理一体化框架并分析相关性能；系统调度方面，设计面向不同应用需求的多工作模式，研究不同模式下的具有资源意识的节点调度策略，尤其针对资源严重受限情况下，提供性能折中的权衡策略。此外，针对高精度的目标监测需求，根据基于数据压缩的目标直接定位思想，研究适合新体系特点的目标跟踪框架及算法。在理论研究基础上，开发新体系下的原型系统，解决传感器节点高精度同步与自定位等支撑技术难题，并验证算法及模型的性能。此课题为大规模实时无线传感器网络突破实时性与资源限制提供了解决框架并进行实验验证，为其实用化提供了新途径。

面向高效实时的目标监测，本课题在阵列技术思想的启发下，以最新的压缩采样技术为支持，提出了一种新型的阵列化传感器网络体系架构，重点研究资源受限的实时传输与目标监测。.信息实时传输方面，根据目标监测应用中负载动态变化的规律，本课题提出了一系列采用多信道技术、跳频技术及动态时隙分配技术的负载自适应协议，以保证低负载时系统的低功耗及突发负载下快速高效的实时数据传输。此外，考虑到声信号等存在稀疏性，本课题将信号的有损传输模拟为随机压缩采样过程，以保证稀疏信号的实时高效传输；.实时高精度目标监测方面，本课题首先提出了阵列压缩采样-DoA联合稀疏重构方法以减少阵列处理冗余及误差传递。同时面向低功耗无线目标监控场景下可能存在的资源极度受限的无线传感器场景，我们提出一种基于单比特深度量化的DoA估计方法，进一步减小了对系统储存、通信资源的消耗。面向网络化系统存在的降低计算能力需求，我们提出一种极低计算复杂度的快速三维目标实现，从而在低功耗、低成本本地节点实现复杂的三维目标DoA估计。此外，本课题还提出了面向纯方位角目标定位的位置受惩极大似然估计器及基于辅助变量的节点自定位闭环算法。.在理论研究基础上，本课题开发了新体系下的原型系统，解决传感器节点高精度同步与自定位等支撑技术难题，并验证算法及模型的性能。此课题为大规模实时无线传感器网络突破实时性与资源限制提供了解决框架并进行实验验证，为其实用化提供了新途径。