面向跨层优化的三维时空数据收集方法

Three-dimensional spatial-temporal data collection is one of the key issues in the study of spatial-temporal information perception in the Internet of Things (IoT). However, the research challenges and computational complexity are much higher than that of two-dimensional problems. Thus, the applicability of the research results obtained by directly extending two-dimensional conclusions to the three-dimensional scenario or conventionalizing three-dimensional problems to two-dimensional scenario is limited. In our research project, we mainly focus on the basic theory and key technology research of complex three-dimensional spatial-temporal data collection. Firstly, the Fuzzy Analytic Hierarchy Process (FAHP) is applied to judge the importance of three-dimensional node perceptual data. Based on the importance of node perceptual data, the sparse key information node set of the network is constructed to dynamically reconstruct the network information. Thus, the unnecessary node data transmission will be reduced and the network redundant data transmission quantity will be decreased. Secondly, the collision-tolerant scheduling is proposed to ensure the priority of data operation for key nodes to reduce data transmission delay in data collection and routing process. The node scheduling with low-latency in three-dimensional environment with complex interference will be realized. Finally, a lightweight and self-adaptive spatial-temporal data aggregation routing strategy integrating energy, delay and reliability is proposed to realize cross-layer optimization of three-dimensional spatial-temporal data collection under low hardware system configuration. Studies of the project are able to provide complex three-dimensional spatial-temporal data collection with fresh thoughts, and support theoretical guidance to the further development of IoT spatial-temporal information perception and spatial sensor network monitoring. Moreover, the independent innovation of key technologies of IoT spatial-temporal big data perception in China will be promoted.

三维时空数据收集是物联网时空信息感知研究的关键内容之一，但其难度和计算复杂性远高于二维平面，而直接推广二维结论到三维或者规约三维问题到二维平面所得研究成果的适用性受到限制。本项目以复杂三维时空数据收集的基础理论和关键技术研究作为主攻方向，提出利用模糊层次分析法解决三维感知节点数据重要性判定问题，基于节点感知数据重要性构建网络的稀疏关键信息节点集，减少不必要节点数据传输，降低网络冗余数据传输量；提出数据收集与路由过程冲突情况下的关键节点优先数据操作以降低数据传输时延的调度方法，实现三维复杂干扰环境下的低时延节点调度；提出综合能量、时延和可靠性的轻量级自适应时空数据聚合路由策略，实现低硬件系统配置下三维时空数据收集的跨层优化。本项目的研究为复杂三维时空数据收集提供新思路，且为物联网时空信息感知和空间立体传感网监测的进一步发展提供理论支撑，也将促进我国物联网感知时空大数据关键技术自主创新。

针对嵌入式传感设备部署的大规模性和能量有限性，监测物理对象和环境的复杂性、动态性和无线通信的不可靠性，及基于传感设备的三维时空数据的收集面临数据量暴涨、传输干扰复杂、边缘网络开放性带来的能耗、时延和可靠性问题，课题围绕“复杂三维时空数据收集的基础理论和关键技术研究的主攻方向，层次化数据收集研究思路”，研究网络信息动态重构的稀疏节点集构建、复杂干扰环境下节点调度、轻量级时空聚合数据路由策略。在稀疏节点集构建方面，提出了基于冗余节点判别的稀疏工作节点集合构建策略；且通过研究时空数据收集中节点的可信度，提出一种主动可验证信任评估方法识别感知设备的可信性，从而保证低成本可靠的边缘计算数据采集，并通过众包思想招募多个移动终端对感知设备进行可信度评价以构建稀疏可信感知设备节点集合，实现可信低冗余数据收集。针对三维感知数据收集，提出了一种三维感知网络的节点负载和传输延迟分析模型。在采用最短路径路由的情况下，利用微元思想给出了网络中每个节点的数据负载、能量消耗和E2E延迟分析模型。并基于此，提出了一种能量感知的数据传输和调度方案，在保证传输可靠性的前提下，实现了E2E时延与网络生存时间的折衷优化。基于车辆边缘网络应用场景，提出一种车辆边缘计算均衡负载解决方案。通过将资源受限的监控与数据采集SCADA系统中的应用程序卸载到轻量级和无处不在的车辆边缘计算VEC服务器上，实现任务调度和负载均衡。针对时空数据收集中的不规则网络拓扑，提出了基于测地线距离的轻量级数据传输与路由策略。该策略针对不规则部署的感知网络，通过基于测地线距离的感知设备选取，确保不规则网络环境下数据收集中数据转发的方向朝着汇聚节点的位置进行，避免传输空洞，提高传输效率。