Fusion Estimation For Networked Multi-Sensor Systems With Communication Constraints

With the rapid development of3C (Communication, Computer, Control) technology and the emergence of new type sensors, a variety of multi-source information systems for complex application spring up which leads to the urgent need for the development of multi-sensor information fusion (MIF) techniques. As an important branch of multi-sensor information fusion, the multi-sensor information fusion estimation plays an important role in many military and civilian fields. Introducing communication network into network multi-sensor fusion systems (NMFSs) can solve the difficulties in the traditional fusion system with wiring complexity, poor scalability, high maintenance cost, etc. However, the introduction of the communication network brings new chanllenges, including the energy constraints and the media access constraints. Based on the approach for periodic system, optimal estimation theory and the optimal fusion technique, this thesis investigates these two chanlleges. The main work of the thesis is summarized as follows:1) The H∞estimation problem is investigated for NMFSs with energy constraints. An energy-efficient transmission strategy is proposed by slowing the measurements transmission rate from the sensors to the fusion center, and the resulting estimation system is described as a linear discrete-time periodic system model. Then, by using the Lyapunov stability theory and Linear matrix inequality (LMI) technique, and the approach for periodic system, a sufficient condition is presented for the estimation system to be asymptotically stable and ensure a prescribed H∞performance. Moreover, a design method for the H∞fusion estimator is also presented. Finally, a numerical example is given to demonstrate the effectiveness of the proposed approach.2) The Kalman estimation problem is investigated for NMFSs with media access constraints. A novel periodic transmission strategy is proposed, and each local estimation system is modeled as a discrete periodic subsystem. According to the latest measurement information update time, each subsystem chooses the appropriate Kalman estimator (filter or predictor) to obtain the local optimal estimate, and the fused estimates are produced by applying the optimal fusion criterion weighted by matrices. Finally, a target tracking example is given to show the effectiveness of the proposed method.3) At last, a summary of the full thesis and some prospects for future research are made.