基于众包和群智计算的室内无线定位理论和方法

In the recent years, with the ever-increasing demand for indoor mobile location-based services (MLBS), the indoor localization technology with sensors integrated in smart terminals has become a hot topic. To cope with the heavy training overhead, inconvenient deployment and maintenance of the existing indoor localization technologies, this project will focus on the fundamental theory and methods of indoor wireless localization based on crowdsourcing and crowd computing. The research topics in this project cover most key parts of the indoor wireless localization based on crowdsourcing. The main research contents includes hybrid sensor signals pattern mining of indoor landmarks, dead reckoning with high accuracy, construction of indoor floor plan, automatic location labeling of training fingerprints, multi-mode fusion localization algorithm, low-power positioning technology, etc. We will adopt data mining, pattern recognition, clustering, manifold learning, graph theory, Bayesian inference and other methods to explore the above problems. After the implementation of this project, a series of innovative research results about crowdsourcing wireless localization will be achieved, and a zero-effort, high accurate and full-coverage localization algorithm will be proposed, which is also adaptive to the indoor environment change. This project helps to lay a theory foundation of large-scale application of indoor localization technology for smart terminals. The expected results include 12-15 papers,6-8 patents,2 Ph.D and at least 10 masters candidates training.

近年来，室内位置服务需求日益迫切，基于移动智能终端集成传感器的室内定位技术成为研究热点。本项目针对现有室内定位技术存在训练开销过大、部署和维护困难等问题，开展基于众包和群智计算的室内无线定位基础理论和方法研究。项目研究内容涉及基于众包无线定位技术的多个核心环节，主要包括室内地理特征点混合传感信号特征模式发掘、用户室内移动轨迹准确推断、室内平面布置图挖掘、训练指纹地理位置自动标记、多模融合高精度定位算法和低功耗技术等关键科学问题。项目拟采用数据挖掘、模式识别、聚类、流形学习、图理论以及贝叶斯推断等方法对上述问题进行理论探索。通过本项目的实施，将在基于众包的无线定位理论和方法上取得系列创新成果，并探索出适合于室内环境的零开销、高精度、全覆盖定位方法，为面向移动智能终端的室内定位技术的大规模应用奠定理论基础。预期发表高质量学术论文12-15篇，申请发明专利6-8项，培养博士2名，硕士10名以上。

近年来，室内位置服务需求日益迫切，基于移动智能终端集成传感器的室内定位技术成为研究热点。本项目针对现有室内定位技术存在训练开销过大、部署和维护困难等问题，开展基于众包和群智计算的室内无线定位基础理论和方法研究。项目研究内容涉及基于众包无线定位技术的多个核心环节，主要包括室内地理特征点混合传感信号特征模式发掘、用户室内移动轨迹准确推断、室内平面布置图挖掘、训练指纹地理位置自动标记、多模融合高精度定位算法、用户个性对众包影响和低功耗技术等关键科学问题。项目采用数据挖掘、模式识别、聚类、流形学习、图理论以及贝叶斯推断等方法对上述问题进行理论探索。通过本项目的实施，在基于众包的无线定位理论和方法上取得了系列创新成果，探索出了适合于室内环境的零开销、高精度、全覆盖定位方法，为面向移动智能终端的室内定位技术的大规模应用奠定了良好理论基础。预期发表高质量学术论文34篇（其中SCI索引论文6篇），申请发明专利10项（其中获得授权7项），参与国家实时定位标准制定4项，获得奖励3项，培养博士2名，硕士28名。