Research On UDR Based Urban Sensing And Application

Urban sensing is one of the most popular research areas, and shows great importance for urban planning, community service, and public resource allocation. With the rapid development of global urbanization in recent years, population explosion, resource shortage, environment degradation, and traffic congestion have brought greater challenges for urban issues than ever before. Since human act as the subject of city, the key factors for meeting above challenges lie in human behavior analysis. Based on the characteristics of human behavior and resource consumption, urban managers can obtain strategies for management and optimization. A well accepted framework for building smart city is to utilize information and communication technology to sense and analyze the key indicators for urban operation, then management and optimization are conducted based on these key indicators.Benefit from the rise of mobile internet, abundant data containing spatial-temporal tags and describing personal behavior make it possible for a fine granularity analysis towards urban space. However, mobile communication system requires increasing energy, and bandwidth resource shows greater shortage. Serving as a typical data source, usage detailed records (UDRs) not only include plentiful and variety of spatial-temporal information with small intervals, but also contain detailed mobile communication resource consumption for every user. Moreover, UDRs provide new insight for urban sensing, and promote the emergence of mobile communication system optimization method under current technical background.Therefore, this dissertation conduct urban sensing task based on the UDRs obtained from Guizhou province. In particular, this dissertation proposes a novel network model for community detection in urban sensing, and presents an efficient community detection method utilizing the novel network model. Moreover, this dissertation applies the proposed community detection method to optimize energy consumption and communication efficiency in mobile communication system. The main achievements of this research are presented as follows:(1) Based on usage detailed records, a novel shared network is proposed for urban sensing. This dissertation project large number of user's mobile movements into urban space and sense its structure.(2) This dissertation explores the community detection method based share network model. In particular, we study the spatial-temporal behavior hidden in usage detailed records and propose a novel community detection method based uncertainty and redundancy, which can be applied in urban planning and fine management.(3) A novel energy saving scheme specified for mobile communication system is proposed for urban sensing. Urban management is eventually realized in infrastructure. Base stations are widely distributed in mobile communication system and consume huge energy. This dissertation proposes a saving scheme for mobile communication system based traffic consumption in usage detailed records. Simulation results show that our proposed scheme has great advantage over traditional methods.(4) This dissertation provides a traffic optimization scheme for mobile communication network in urban sensing. Since the distribution of base station and spatial-temporal behaviors of mobile users have great impact on communication resource usage efficiency, this dissertation conducts a traffic optimization scheme based traffic sensing, which dynamically allocate bandwidth resource in core network according to the requirements of base stations, to boost bandwidth usage efficiency. Validated by data in physical world, simulation results show that the proposed scheme has advantages in efficiency.This dissertation is an exploratory research to urban sensing based on usage detailed records. We sense the urban space structure through network traffic, and project users' movement to base station network. Based on the sensed results, we propose both energy saving scheme and traffic optimization scheme for infrastructure management in mobile communication system. The network model, community detection method, and management scheme presented in this dissertation are helpful in understanding the structure of urban space, planning urban areas, and realization of green communication. They also provide new insight for urban sensing, and promote the development of urban computing.