Research On Key Techniques Of Distributed Sensor System Based On WSN

Wireless sensor network is one of the most important technologies in the 21st century,and its extensive applications have greatly improved the ability of human cognition and transformation of the world.Wireless sensor networks have a distributed nature.Sensor node is the basic unit of a network,its calculation,storage and communication capacity is extremely limited.How to rationally allocate node resources,to achieve optimal configuration and constitute a powerful network,is the key to sensor management.In this paper,the problems of collaborative estimation and sensor scheduling will be studied based on the application of remote network monitoring system.We will discuss how to use local communication of sensors to optimize the measurement fusion and achieve local state estimation;and how to schedule data transmissions so that the remote state estimates meet given accuracy requirements.Because there are many monitoring processes in the remote monitoring system,centralized processing approach involves mass data transmission and affects the overall efficiency of the system.In this paper,the problem of distributed estimation for relative sensing networks is studied to eliminate redundant information and reduce data transmission.In the network,all nodes are able to measure the relative states between themselves and their neighbors.Anchor nodes can additionally measure their own absolute states.A bi-directional graph is used to model the network topology in this paper.By constructing the optimal estimator,it is discussed that a connected graph is necessary for global state estimation.Two Kalman like suboptimal estimators and one average consensus suboptimal estimator are designed.Thus,the local estimation of processes' global states is realized.In order to avoid communication conflicts and reduce the energy consumption of the wireless sensor network,the problem of scheduling parallel Kalman filters is studied in this paper.The goal is to plan sensor transmissions in the single-channel communication restrictions,so that all state estimates of processes in the remote center meet the given accuracy in all times.In this paper,each process corresponds to a sensor and a remote estimator.A real-time deadline of transmission is defined and calculated.The problem of sensor scheduling based on covariance is thus transformed into a task planning problem.Then the off-line scheduling method based on transmission periods and the on-line adjustment method based on sliding windows are designed,which can adapt to transmission failures and threshold changes.When there are a large number of processes in the system,using a single-channel only can not guarantee timely data transmission.Therefore,an adaptive multi-channel transmission scheduling is proposed to reduce the channel occupancy while preserving the estimation accuracy.In this paper,a squeeze method that can accurately calculate process transmission cycles is presented.In order to meet multi-channel transmissions,an off-line schedule method based on the Buffer Scheme and a corresponding on-line adjustment method are proposed.The multi-channel algorithm simplifies the computation of the adjustments,improves the idle channel utilization efficiency,shortens the adjustment time after the system changes,and is more adaptable to distributed networks.The algorithms proposed in this paper will be analyzed in detail in the corresponding chapters.Several simulations are given in each chapter to illustrate the results.