面向低占空比无线传感器网络的跨层QoS保障协议研究

In duty-cycled wireless sensor networks, the performance of QoS depends on the key factors in the different layers of the protocol stack, which makes it impossible for the single-layer optimization methods to provide the QoS guarantee. This project adopts the analysis of cross-layer theory, the interrelationship between the QoS metrics and the key factors are derived for supporting the modeling of the related key performance indicators. By building the analysis model of congestion based on the throughput QoS, an enhanced congestion control protocol is proposed with the dynamic re-routing mechanism. With the analysis model of the single-hop delay based on the QoS delay, an end-to-end delay control protocol is proposed by adaptively adjusting and optimizing the duty-cycling. With the improved detection model of the QoS packet loss, via the analysis and optimization of the threshold value, a detection protocol for malicious packet loss is proposed for satisfying the required false negative rate and false positive rate. This project aims to explore the new solution for the cross-layer protocols design under the asynchronous sleep schedule mode, which can improve the QoS metrics, including the performance in throughput, delay and packet loss detection. The proposed protocols in this project are designed under the unified cross-layer structure of the component-based development, which indicates the organic unity of the QoS metrics. This work will provide new ideas and new methods of designing the reliable, real-time and security cross-layer protocols for the duty-cycled wireless sensor networks.

低占空比无线传感器网络的QoS性能取决于协议栈各层之间的多个关键因素的影响，使得基于单层的优化方法无法提供QoS保障。本项目在异步睡眠调模式下，利用跨层理论进行分析，解析推导QoS性能指标与关键因素之间的相互关系；建立基于QoS吞吐量的拥塞度分析模型，利用动态重路由机制，提出增强型拥塞控制协议；建立基于QoS时延的单跳时延分析模型，通过自适应调整和优化占空比，提出端到端时延控制协议；改进基于QoS丢包分析的恶意丢包攻击检测模型，通过阈值分析和优化，提出满足一定误报率和漏报率的恶意丢包攻击检测协议。本项目旨在探索满足QoS吞吐量、时延和丢包检测的跨层协议设计的新解决方案，所提出的协议采用统一的基于组件的跨层设计框架，体现了QoS性能指标之间的有机统一，为低占空比无线传感器网络的应用提供具有可靠性、实时性和安全性的跨层协议设计新思路、新方法。

低占空比无线传感器网络的QoS性能取决于协议栈各层之间的多个关键因素的影响，使得基于单层的优化方法无法提供QoS保障。本项目在异步睡眠调模式下，利用跨层理论进行分析，解析推导QoS性能指标与关键因素之间的相互关系，主要研究内容包括：（1）基于低功耗监听模式和队列模型，将增强型拥塞控制问题归约为链路拥塞和队列拥塞的判定问题求解，提出了满足QoS吞吐量的拥塞度检测方法，用于动态重路由机制的路由选择度量，实现了增强型拥塞控制的QoS感知路由协议；（2）基于跨层设计思想和单跳时延分析模型，将端到端时延控制问题归约为带约束的目标优化问题求解，提出了满足QoS时延的保障方法，用于动态调整睡眠调度时间间隔，实现了满足端到端时延需求的跨层自适应占空比的控制协议。（3）基于Watchdog的丢包行为检测和链路质量估计，将恶意丢包检测问题归约为阈值最优化问题求解，提出了满足QoS丢包的恶意丢包攻击检测方法，用于满足一定的误报率和漏报率需求，实现了QoS感知的安全路由协议。（4）为了验证本项目提出的相关算法和协议，利用TinyOS基于组件的协议体系结构，满足节点间通知报文广播、邻居发现和邻居信息交互等需求，引入延迟广播转发策略，提出了一种面向异步睡眠调度的时延受限和能量均衡的多跳广播协议。另外，瞄准当前区块链研究热点，结合本项目相关研究基础和物联网关键技术，提出了基于节点行为特征分析的数据传输模式，并进一步研究基于拜占庭容错协议一致性问题，实现能够根据区块链的结构评估和检测安全性的方法。本项目为低占空比无线传感器网络提供满足QoS吞吐量、时延和丢包检测的跨层协议设计的新解决方案。本项目组成员明确了区块链技术的物联网研究面临的关键问题，研究满足安全性、可扩展性、TPS性能以及资源有效性的共识机制，为工业实践领域对区块系统的QoS感知的物联网应用提供创新性解决方案，在区块链技术落地和区块链系统性能优化方面做出新尝试。