基于能量收割技术的网络化非线性随机系统的滤波与控制

In this project, the filtering and control problems are investigated for several classes of networked nonlinear stochastic systems with energy harvesting sensors. By analyzing the energy harvesting mechanism thoroughly, the dynamically evolution rule of the energy stored in battery is revealed and then its mathematical model is established. Then, the dynamic of the energy stored in battery is analyzed quantitatively and the measurement output model using energy harvesting technology is established. By taking into account the natures of randomness, nonlinearity and the feature of energy harvesting, the filtering problems are investigated for nonlinear stochastic system, sensor networks and complex dynamical networks with network-induced incomplete information and the effective filtering schemes are provided. Based on them, in the case that the system states are unknown, the dynamic output feedback control problems are studied for nonlinear stochastic systems and complex networks with energy harvesting sensors. The control approach and the synchronization control approach are obtained for nonlinear stochastic systems and complex networks, respectively. Finally, the results derived are demonstrated in an experimental platform established in the Networked Robot Control Laboratory. The research results obtained in this project are of a great significance both in theory and application prospect.

本项目主要研究几类基于能量收割技术的非线性网络化随机系统的滤波与控制问题。全面分析能量收割技术的原理，揭示带有能量收割装置的电池存储能量的动态变化规律并为其建立数学模型，然后对建立的电池能量的动力学模型进行定量分析，从而建立基于能量收割技术的测量数据传输模型；针对系统的随机、非线性特性以及能量收割技术的特点，综合考虑网络化环境下多种测量信息不完全现象，分别研究非线性随机系统，传感器网络和复杂动态网络基于能量收割技术的滤波问题，并提出合理的滤波方案；在此基础上，针对状态未知情形，进一步考虑非线性随机系统和复杂动态网络基于能量收割技术的动态输出反馈控制问题，提出非线性随机系统的控制和复杂网络的同步控制方法；最后，本项目得到的理论成果将在网络化机器人控制实验室中进行实验验证。本项目成果具有重要的理论意义和广泛的应用价值。

本项目主要研究了几类基于能量收割技术的非线性网络化随机系统的滤波与控制问题。全面分析了能量收割技术的原理，揭示了带有能量收割装置的电池存储能量的动态变化规律并为其建立了数学模型，然后对建立的电池能量的动力学模型进行了定量分析，从而建立了基于能量收割技术的测量数据传输模型；针对系统的随机、非线性特性以及能量收割技术的特点，综合考虑网络化环境下如传感器饱和、欺骗攻击、通信协议、信号衰落、中继传输机制、采样等多种测量信息不完全现象，分别研究了非线性随机系统，传感器网络和复杂动态网络（神经网络、时空系统）的同步与状态估计等问题。研究过程中针对遇到的优化问题，研究并提出了新型智能优化算法。基于所得理论成果，在机器人系统、电动车和电力系统中进行了应用研究，并在相应的平台上进行了验证。在本项目资助下，申请发明专利4项；培养硕士研究生5名，博士研究生4名；发表期刊会议论文51篇，其中SCI论文43篇，有7篇发表在控制学科顶级期刊IEEE Transactions on Automatic Control和Automatica上，其他IEEE汇刊论文8篇。