局部配置无线对等感知采集网络的度聚集混合互编码策略研究

Peer-to-peer wireless sensor network is often used in data monitoring and collection applications in disaster or other special environment. For the special requirements of regional network data survivability and the efficiency of data recovery, this project intend to carry out research on reliable data recovery and optimized storage problem for peer-to-peer wireless sensor network with both data recovery protocol and storage strategy. In order to improve the efficiency of the growth code, lightweight-routing trend converge tree is introduced and the local configuration network model is constructed. The discrete distribution mode is applied to enhance the data mobility, and the degree conversion strategy is optimized. Multi-point two-level monitoring cache-filter mechanism is designed to relieve the delay effect in the terminal period of data collection. On the other hand, we optimize the system performance by adjusting storage strategy. Degree priority distribution framework is introduced with edge-degree aggregation and level-by-level activation, so that the cliff effect in the early period of data collection can be relieved, and the collection overhead can be reduced and the survival rate of sensor data can be improved. Finally, the decoding and acquisition characteristics of multiple generation data collection framework is observed. Optimizing the collection cycle conversion strategy and designing hybrid mutual coding strategy with the data dependence within snapshots, the system acquires the mutual decoding and protection ability to reduce the storage costs. The model, strategy mechanism and the algorithm protocol can be applied to provide theoretical models and solutions for reliable data collection of peer-to-peer wireless sensor network, and will have a positive role in this field.

无线对等感知网络常应用于恶劣或特殊环境数据监测与采集，针对该类网络对区域数据存活性及采集效率的特殊要求，本项目拟从采集协议与存储策略两方面协同开展无线对等感知网络数据可靠性回收及优化存储问题研究。引入轻量路由趋向汇聚树，构造局部配置网络模型，以离散分发增强数据流动，优化度转换时间序列，从而提高增量码类协议效率；以多点二级监听缓存及预过滤，强化对数据收集末端延迟效应的消解。另一方面借助存储策略优化系统性能，引入“度”依边缘聚集并按层次激发的度优先级分布框架，强化对数据收集早期陡壁效应的消解，降低收集时间开销，加强存储数据的存活率。最后研究多代采集框架下的解码及采集特性，优化采集周期转换，利用数据依赖设计混合互编码策略，获取隔代数据的互解码及保护能力并降低存储开销。本研究建立的模型、策略机制及算法协议，将为无线对等感知网络的可靠数据收集提供理论模型和解决方案，对该领域产生积极的推动作用。

本项目在基于无线对等感知网络的数据采集系统中，针对现有协议采集效率不高、数据存活率偏低的状况，开展了基于局部配置信息的编码数据采集策略研究、边缘度聚集的无线对等感知网络编码可靠性存储策略研究、多代数据采集框架下数据依赖的混合互编码策略研究等一系列理论探索和系统开发工作。突破传统方法在收集策略及网络构造上的同质化和单纯化特征，探寻采集+存储协同优化的方向，为对等网络体系设定略微复杂但更有利于数据分发及汇聚的配置和行为模式，并扩展数据编码的维度,使之适用并能利用多代连续数据采集模型的时空约束特性，从而最大限度地提高采集模型的性能及数据可靠性。并将此项目部分技术手段应用于舟山渔船信息采集与分发系统中，为行业实践提供了高效可靠的数据分发及同步支撑。