基于多源异构数据融合的一体化室内位置服务技术研究

Location-based service is one of the most crucial techniques in networking. It has brought many practically innovative applications to various fields, such as transportation, logistics, people tracking and emergency service. For instance, visitors could reach each exhibition room with the floor plan of museum. Shop owners could customize floor plans to smoothly guide potential customers to their shops. Previous approaches collect location-related information mainly from wireless communication network of telecom operators or external localization, including WiFi and GPS. Because of inherent limitations, existing mature location-based services are appropriate for outdoor settings. However, there is a great demand for indoor location-based services, such as indoor map reconstruction for disaster rescue and emergency, localization and navigation in parking lot, shopping mall and airport. These problems are valuable because an accurate indoor floorplan provides better location-based services and facilitates information sharing, optimal navigation, tracking and location-based event identification. Therefore, how to efficiently provide various indoor location-based services remains to be an open but critical problem..To solve above important problem, this project will propose new efficient approaches to providing indoor location-based service and seamless connection between outdoor and indoor service. To this end, we will provide effective methods and integrated systems to address three new sub-problems, indoor map reconstruction, indoor map update and indoor localization. These three problems are relevant - for instance, a reconstructed indoor map can assist the updating of indoor map; an up-to date indoor map can improve the accuracy of indoor localization. Our approaches can make contributions as follows: reconstruct indoor map with abundant labels including elevators, escalators, stairs, names and categories of facilities, and stores; identify outdated data in the current indoor map and then update them with new collected data; accurately localize users in indoor environment with the up-to-date indoor map. The explicit novelty of our project is two-fold by reconstructing indoor map from sensing data and images acquired only from smartphones, and providing accurate indoor localization and navigation without extra devices. By solving critical problems in indoor location-based service, our project will generate integrated system designs and application approaches, with a long-term impact on multiple areas of research and practical application, such as security, social behavior, ergonomics and design of buildings and indoor facilities, logistics and functional optimization, as well as in communications and cloud computing for smart cities.

基于位置的服务是最为重要的技术之一，它已被广泛应用于交通、物流、人口跟踪、急救服务等领域。目前市场上对于室内位置服务的需求是非常强烈的，如抢险救灾时获取的室内地图信息，停车场、商场、机场等场所的定位和导引。本项目基于以上的问题，提出了构建全方位一体化室内位置服务的解决方案，实现室内外位置服务无缝衔接的良好用户体验。本项目针对室内地标复杂和传感器失准场景下低成本地图构建问题，多元室内环境变化及噪声下有效准确地图更新问题，以及单一室内定位方法失效下，进行多种定位方法融合的定位问题，进行研究和提出了有效的解决办法，并设计了整体系统方案。本项目的创新点是仅利用智能手机传感器及照片获取原始多源异构数据，构建与更新地图，并在此基础上提供精准定位服务，不使用额外的设备，可用性高。本项目的特点是不仅解决了提供室内位置服务的关键问题，还提出了系统设计和应用方案，无论从科研还是从实际应用中都有非常重要的意义。

本项目针对室内位置服务不精确的问题，提出了以室内地图为基础，进行室内地图构建、室内地图更新、室内精准定位等关键技术的设计，提供室内位置服务一体化的解决方案。在地图构建阶段，我们提出了基于智能手机的传感器数据，构建室内地图的高效方法。在地图更新阶段，我们利用街景和感知移动用户众包的数据，提出了数字地图的自动智能更新系统。在精确定位阶段，我们设计基于WiFi，RFID 和航迹推算的定位方法，以及利用基于图像特征的图像定位算法，并综合多种定位方法提出了适用于室内的精确定位算法。.本项目的成果还包括高效算法的设计，应用于众包收据的获取及隐私保护，手机数据的高效处理和室外导航设计，室内导航算法和系统的设计，室内手机数据的安全，隐私保护和分析。利用本项目的成果，我们还成功申请了一项国家专利，对物联网设备的使用安全性设计了新的安全授权方法。此项目的意义是我们的方法对移动数据的获取，处理，比传统方法具有更好的室内定位效果，能提供更精确的室内位置服务，并从数据安全和隐私的角度给出了解决方案。此项目的成果还可以应用于不同的移动计算场景，比如物联网。.项目的成果发布在计算机领域的顶级会议和顶级期刊上，包括9篇会议文章和15篇期刊论文。顶级的会议论文包括IEEE IWQoS2020, IEEE SECON2019, IEEE INFOCOM2018, ACM UBICOMP2018, ACM ASIACCS2018。顶级的期刊包括IEEE TMC 2021-1, IEEE TMC 2021-2, IEEE IoTJ 2021, IEEE Wireless Communications 2020, IEEE TSMC 2020, IEEE J-SAC 2019, IEEE TMC 2019-1, IEEE TMC 2019-2, IEEE TMC 2019-3, IEEE IoTJ 2018, IEEE Network 2018-1, IEEE Network 2018-2, IEEE T-ASE2018, IEEE TVT2018。