基于多模态信息的自校准高解析度被动式室内定位研究

Passive indoor localization relies on no explicit cooperation of mobile devices and has wide range of applications in smart cities, and it has become one of the most active research areas in the mobile computing community recently. Due to the devices diversity, limited information, and sparsity of transmissions, passive indoor localization suffers from low accuracy, difficult deployment, and poor robustness in practice. To the best of our knowledge, this project is the first systematic study on the super-resolution, high robustness passive indoor localization in challenging indoor environments. In this project, we will research the multimodal super-resolution algorithms to achieve centimeter-level accuracy in indoor environments with severe multi-path effect. To reduce calibration and maintenance effort, we further investigate the self-calibrating mechanisms of passive indoor localization. Next, passive transmission prompting mechanisms and continuous tracking mechanisms will be studied to support tracking services. This project builds an efficient framework on passive indoor localization from localization algorithm, self-calibrating mechanisms, transmission prompting mechanisms to the tracking algorithms. We believe this project will have both important theoretical and practical significance to passive indoor localization research.

被动式室内定位技术消除了传统室内定位需移动设备主动配合的依赖，在智慧城市建设中具有非常广阔的应用前景，已成为当前移动计算领域的一个重要研究方向。由于被动式室内定位具有设备多样化、有效信息少、传输稀疏性等特点，目前存在精度低、部署难、鲁棒性差等问题，尚无相关研究能很好地解决这一问题。本项目构建理论分析框架并研究若干具有创新性的方案，研究多模态高解析度被动式室内定位算法，在存在复杂多路干扰的室内环境中实现厘米级高精度定位；研究多模态被动式室内定位系统的自校准、自维护机制，降低系统部署及维护开销；研究被动式传输触发机制及被动式连续追踪机制，实现高鲁棒性被动式连续追踪。本课题将形成被动式室内定位从定位算法、校准机制、触发机制到追踪机制的一套有效的理论方法，为被动式室内定位的长期大规模实际应用奠定理论与实践基础，促进智慧城市建设的创新与繁荣。

本项目针对被动式室内定位存在的精度低、部署难、鲁棒性差等问题，构建理论分析框架并研究若干具有创新性的方案，研究多模态高解析度被动式室内定位算法、多模态被动式室内定位系统的自校准、自维护机制，以及被动式定位连续追踪机制，实现高鲁棒性被动式连续追踪。针对被动式定位精度问题，项目提出了基于PSO优化的高精度定位方法，以及基于Spectrum Matching的细粒度相对定位方案，提升了被动式定位精度；针对定位系统部署及维护难问题，项目提出了基于路径映射的自校准机制、以及基于地标智能是别的语义地图生成机制，从而显著提升了定位系统的自校准能力；针对被动式定位鲁棒性及连续追踪问题，项目提出并验证了动态环境下基于连续追踪定位的线下社交网络重建系统，实现了被动式定位连续追踪应用。项目构建了被动式定位定位算法、自校准机制及连续追踪机制的理论及应用框架，为未来被动式定位的进一步应用发展奠定了基础。