Distributed Multi-objective Filtering Over Wireless Sensor Networks

Wireless sensor networks have been a key technology for the Internet of Things(IOT).It is of important application value on military defense,industrial control,environmental monitoring,fine agriculture and smart home.With the development of wireless sensor networks,filtering algorithms based on the wireless sensor networks have also extensively studied.In this thesis,non-zero Nash game is used to address the problem of distributed multi-objective filter design over lossy wireless sensor networks for systems subject to both stochastic noises and modeling errors.Based on stabilizing solutions in the mean square sense,Nash equilibrium strategies,consisting of the optimal filter gains and the corresponding worst-case disturbance signals,are analytically conducted.Moreover,to reduce the disagreement among different local estimates,distributed consensus robust optimal filters are established for each node.The stability range of consensus parameter is given by a modified algebraic Riccati inequality,together with the complexity of the network's topology.Specifically,the contexts are divided into:The first part: the definition of the distributed robust optimal filters with packet loss is formulated.Based on the TCK-like packet loss protocol,the structures of the distributed robust optimal filters are given.To tackle both stochastic noises and modeling errors,two cost functions are used for describing the robustness and the optimality associated with the filter system.Then,the problem of distributed robust optimal filters design is defined within the formwork of Nash game.The second part: this part is about the design of the distributed robust optimal filters with packet loss.The stabilizing solutions in the mean square sense are established for a set of cross-coupled modified algebraic Riccati equations.Based on the stabilizing solutions,Nash equilibrium strategies,consisting of optimal filter gains and the corresponding worst-case disturbance signals,are further analytically conducted.The third part: this part is dedicated to the design of the distributed consensus robust optimal filters.To reduce the disagreement among different local estimates,the distributed consensus robust optimal filters are provided to achieve additional consensus objective.Furthermore,through a modified algebraic Riccati inequality,the stability range of the consensus parameter is deduced by using the complexity of the network's topology.The fourth part: in this part,to show the effectiveness of the proposed results we use an example of speed monitoring through wireless sensor network.The experiment platform and the model of a mart car are introduced.Based on an iterative algorithm,the numerical solutions to the cross-coupled modified algebraic Riccati equations are obtained.Furthermore,the distributed robust optimal filter gains and the worst-case signal gains are provided.Then,compared with existing methods,the effectiveness of algorithms is given through simulations and experiments.Finally,the results of this thesis are summarized,and the perspectives of future research directions are also presented.