Research On Coverage And Resource Optimization In Wireless Sensor Networks

With rapid advances in sensors, computing, communications and MEMS technolo-gies, wireless sensor networks (WSNs), integrated with capacities of sensing, information processing and transmission, begin to emerge. Currently WSNs are widely applicable in a large range of fields such as environment surveillance, E-health care, architecture and smart grid, etc. Due to the proliferation of the Internet of thing industry, WSNs are penetrating into our daily lives, changing the way we work and live.Attracted by their tremendous potential and bright blueprint, the academic and in-dustry communities devote huge effort and money to in-depth investigation of WSNs. As one of the fundamental problems in WSNs, coverage has been a hot and challenging re-search topic. Most of existing works aim at addressing the coverage problem in a general way. However, as can be seen from its definition, coverage focuses on the quality of sensing (QoS) about the deployment area and different applications correspond to different require-ments of QoS. Therefore, the solutions derived from the general purpose of network design decline to be conservative when applied to specific applications. Based on this observa-tion and latest existing results, this thesis studies the coverage and resource optimization problem under specific scenarios, with the goal of providing the best solution for specific applications. The main work and contributions are summarized as follows.1) For the efficient capture of stochastic events in static WSNs, a (q,p) sensor periodic scheduling mechanism is proposed, based on which the optimal scheduling schemes (q*,p*) for synchronous, asynchronous and regionally synchronous networks are de-rived, respectively. Moreover, coordinated sleep protocol is designed to further reduce the redundant nodes.2) For the coverage problem in static WSNs when sensory data are correlated, truncat-ed greedy algorithm based on Bayesian inference is proposed, and the approximation ratio of its performance to the optimal one is theoretically analyzed. Due to the decen-tralization nature of WSNs, a distributed algorithm with the same performance as the centralized truncated greedy algorithm is designed.3) For the efficient capture of stochastic events in mobile WSNs, the tradeoff between the sensor mobility and performance of event capture ratio is analyzed. By introducing a practical energy consumption model for sensor movement, a metric for evaluating the efficiency of energy use is proposed. Moreover, the close expressions of the metric under different utility functions are obtained.4) For the cost-effective barrier coverage problem, a periodic sensor scheduling algorith-m is proposed, based on which a centralized sensor coordination patrolling algorithm (SCP) is designed, by jointly exploiting sensor mobility and intruder arrival informa-tion. Two distributed algorithms, S-DSCP and G-DSCP, are obtained.5) For efficient correlated data gathering problem, cross-layer optimization approach is adopted. By employing parallel computing, optimization and Lagrangian dual decom-position method, a distributed algorithm is obtained and its optimality and convergence are analyzed theoretically.In the end, the thesis is concluded and some future research works are discussed.