Multi-channel Mechanism For Optimizing QoM In Wireless Monitoring Networks

With scale expansion and rich applications of wireless networks, the performance guarantee, security and stability of networks face more and more challenges. In wireless networks, multiple sniffers real-time collecting data transmitted of users can realize network fault diagnosis and resource management, which are of great significance on enhancing network performance, safeguarding network security and improving user experience. Due to limited sniffers, how to optimize hardware configuration and software scheduling of all sniffers, to make them cover the most internet users, then collect the most network data, so as to maximize the Quality of Monitoring (QoM) of network, has become one of the hot topics nowadays. This dissertation comprehensively summarizes the theory of optimizing QoM and technology status of wireless monitoring networks, mainly focuses on the optimization methods of sniffer channel allocation/selection in order to improve QoM, and demonstrates the effectiveness of proposed methods through theoretical derivation, simulations and actual experiments.Main research work and innovations of the dissertation lie in:(1) Summarized the concept and research status of wireless monitoring network. Summarized the definition, classification and system framework of wireless monitoring networks; Discussed optimizing QoM methods in wireless monitoring networks, mainly introduced various models and methods of improving QoM by optimizing sniffer channel allocation/selection, and comprehensively analyzed and compared the existing methods from the performance evaluation index system; Discussed optimizing QoM problems in wireless monitoring networks, research idea and arrangements of this dissertation.(2) Proposed a Monte Carlo enhanced Particle Swarm Optimization (PSO) sniffer channel selection algorithm, on the problem of sniffer channel selection in the process of data collection in wireless monitoring networks. A two-dimensional mapping particle coding and the corresponding moving scheme are designed. Monte Carlo method is introduced to revise the solutions, so the particle swarm can quickly converge to the optimum solution or approximate optimal solution. A large number of simulation experimental results show that the Monte Carlo enhanced PSO algorithm is superior to other algorithms, with a higher quality of monitoring QoM, lower computational complexity and faster convergence speed. The actual experimental results also show the effectiveness of the algorithm.(3) Proposed a distributed vertex coloring algorithm based on probability selection, on the problem of multi-channel TDMA time-slot scheduling problem in the process of data collection in wireless monitoring networks. First of all, based on wireless monitoring network topology, build a routing tree to form an interference graph, so as to convert the resource scheduling problem into vertex of interference graph double-coloring problem. The goal is to minimize network traffic conflict; according to the objective function, calculate the probability of vertex selecting color combination, and according to the probability, complete the selection of channel and time-slot. Under the condition of different networks, a series of contrast simulation experiment results show that the algorithm can effectively reduce the number of network conflict, improve network throughput, and reduce network transmission delay and scheduling length, achieving a higher data gathering performance of the wireless monitoring networks.(4) Designed a network user information perception strategy based on sequence learning. As a premise of application of channel selection algorithm, the network topology and user information (working channel, communication probability or weight) must be known. This dissertation presents a user information prediction mechanism based on sequence learning, to help sniffer accurately grasp the user working channels and communication probability or weight in period of entire network information collection, which provides important evidence for channel selection algorithm.(5) Designed a wireless monitoring network channel selection algorithm simulation software platform. Software supports user self-defining network scenarios and node binary chart, integrating and compiling user algorithms, testing algorithm implementation effect, graphically displaying the channel selection process and results of monitoring nodes in wireless networks, evaluating algorithm performance indicators. Based on the simulation software, we test the validity of our proposed algorithms and strategies, and get a series of experimental data, further verified the comprehensive effectiveness and deficiency of the presented methods, and laid a foundation for further related research work.This dissertation studies the optimization of QoM channel allocation/selection mechanisms in wireless monitoring networks; puts forward corresponding algorithms, strategies and a simulation platform; builds a relatively complete theory system. The related achievements will promote the development of wireless monitoring network technology and have an important reference value.