Research On Wireless Sensor/Actuator Networks: Architecture And Algorithms

With the progress and combination of the wireless sensor network and the networked control system, wireless sensor/actuator network (WSAN) is arising recently. As a spe-cific type of the emerging cyber-physical systems, WSAN incorporates a number of ad-vanced techniques such as ubiquitous sensing, self-organization, wireless communication, distributed computing, and automated control. It can be widely applied in industrial automa-tion, manufacturing, smart building, modern agriculture, and intelligent traffic, etc. Being able to improve people's productivity and lives, it has gained high attentions of global re-searchers.Control and networked communication, which are mutually dependent and interact-ing, are two core elements of the WSAN. However, in the literature, most related work focuses on optimizing only one of the two elements, while the joint optimization, which is more meaningful, is still in its exploratory stage. Therefore, further researches are very necessary. On the basis of the latest results, this thesis builds an architecture of WSAN including an evaluation system. Based on this, it studies the problems of protocol optimiza-tion and control algorithms, and also joint optimization of these two elements in WSAN.The main work and contributions are summarized as follows.1) A brief literature review on the development of WSAN and related works is provided.2) An architecture of WSAN is proposed. The requirements and design methodologies of each module are explained in detail. In particular, an evaluation system is proposed and a number of performance indicators are described. Moreover, a simulation platform based on OMNeT++and Matlab for WSAN is designed.3) Protocol optimization methods are proposed which target on state estimation applica-tions. Focusing on the IEEE 802.15.4 standard enabled WSAN, two MAC models for both time-triggered and event-triggered WSAN are established, based on which the optimization frameworks and methods are designed.4) Based on the set-point control applications, three distinct control algorithms are de-signed, i.e., a centralized control algorithm, a decentralized control algorithm and a dis-tributed collaborative control algorithm. Corresponding task scheduling mechanisms for the sensors, actuators and possibly controllers in WSAN, which support the opera-tions of the control algorithms, are also designed in detail. Sufficient conditions for the stability of these algorithms are proved.5) A joint optimization framework between communication and control is proposed. The self-triggered control problem is analyzed as a joint optimization example, and, based on the simulated annealing method, a new control algorithm is designed to address it. The sufficient conditions for the stability of the proposed algorithm is given subse-quently.In the end, the thesis is concluded and some future research directions are depicted.