传感器能量驱动下带宽再分配及量化控制策略研究

Sensor energy constraint and network bandwidth constraint are key issues in wireless sensor networks (WSN). The existing researches are aimed at open-loop system, in which the constraint problems are transformed into optimization solution problem. However, in the closed-loop system represented by industrial WSN, coupled with the synergistic effect of the controller, the establishing process of the optimization problem is quite complicated; Due to the large number of parameters to be determined and nonlinear coupling relationship of which, the optimization problem is difficult to solve; During the operation of the system, because that the impact of different types of sensor failure on the system performance is different, it is more challenging to adjust the existing strategy in real time. For the three problems including modeling difficult, solving difficult and adjusting difficult in the closed-loop system, this project is based on a typical industrial WSN, with data transmission mechanism and networked control theory as means to reveal the synergistic effect of energy constraint and bandwidth constraint, and establish the impact mechanism model of data transmission and control strategy on the convergence domain and convergence time of the closed-loop system; Based on cyclic iterative algorithm, the solution methods of data transmission and quantization control under the bandwidth constraint and energy constraint are established; With the help of fault-tolerant control ideal and switching control theory, the node scheduling, bandwidth redistribution and quantization control means are further constructed when sensors failure to ensure the system performance and save energy simultaneously.

传感器能量受限和网络带宽受限是无线传感器网络（WSN）的关键问题。现有研究多针对开环系统，将受限问题转化为优化求解问题。而在以工业WSN为代表的闭环系统中，针对开环的设计方法不再适用。闭环系统中，加之控制器的协同效应，优化问题建立颇为复杂；由于待定参量繁多且非线性耦合，优化求解较为困难；系统运行过程中，鉴于不同类型传感器失效对系统性能影响的差异性，对策略进行实时调整更具挑战。针对闭环系统中建模难、求解难、调整难三大问题，本项目以一类典型的工业WSN为背景，以数据传输机理和网络化控制理论为手段，揭示能量受限和带宽受限的协同效应，构建数据传输及量化控制对闭环系统收敛域及收敛时间的影响机理模型；基于循环迭代算法，建立带宽约束和能量约束驱动下数据传输及量化控制求解方法；借助容错控制思想及切换控制理论，进一步构建传感器失效时节点调度、带宽再分配及量化控制分层面调整手段，旨在保证系统性能同时节约能量。

针对无线传感器网络（WSN）的能量受限和带宽受限问题，基于事件驱动策略、李雅普诺夫理论、线性矩阵不等式理论、zoom策略等方法，通过引入决策器、估计器、滤波器、虚拟系统等方式，研究了WSN的量化控制、传感器调度等一系列系统分析和设计方法。（1）基于WSN各节点间的数据传输关系，研究WSN的系统建模方法；基于传感器节点采集的数据，设计具有两阶段滤波更新的新型扩展卡尔曼滤波等滤波算法，实现对数据的融合估计。（2）在WSN存在网络带宽约束和传感器能量约束情况下，设计基于事件驱动机制的传感器调度和带宽分配方法，以减小数据传输频率或数据传输数量的方式满足网络带宽约束要求，并实现节约能量的目的。基于量化采样数据、估计误差、采样误差等设计驱动策略，并研究了这些传输机制下WSN的系统估计性能和闭环稳定条件。（3）基于zoom策略、李雅普诺夫理论、矩阵理论等开展网络诱导环境（如带宽受限、数据丢包、传输延时、外部干扰、网络攻击等）下的WSN系统的量化控制问题。重点针对传感器节点模态变化导致WSN呈现切换形式时，设计恰当的量化编码控制策略，保障了不可靠信道传输环境下闭环系统的实用稳定性、李雅普诺夫稳定性、渐近稳定性等性能，实现了系统的良性运行。