近年来，卫星定位与导航服务已经渗透到日常生活的各种场景，被广泛用于导航定位、自动驾驶、无人机、野生动物跟踪等多个新兴领域。然而，在城市高楼等附近，卫星信号会受到多径传输、非视线接收等干扰，会产生几十米甚至上百米的定位误差，这被称为“城市峡谷”效应。目前解决城市峡谷的主要途径是添加额外的传感器、多径检测和3D地图，但仍存在定位误差大、可用场景受限的两个短板。本项目拟借助于多个节点的底层协同工作，来克服城市峡谷效应，实现高精度、高可用性的城市定位。研究内容包括：（1）提出一种卫星原始观测量级的协同定位，实现多节点的异步卫星定位；（2）提出基于众包的网络拓扑构建；（3）评估网络规模、网络动态变化、外部噪声干扰等对协同定位的影响；（4）典型应用场景的原型系统设计和验证。通过理论设计和真实系统验证，项目拟在定位精度、可用性两个维度的综合性能上进行突破，为城市峡谷定位提供有力的技术支撑。

In recent years, satellite positioning and navigation services have penetrated to various scenarios in our daily lives, especially in many emerging areas such as vehicle navigation, autonomous driving, unmanned aerial vehicles, and wild animal tracking. However, satellites signals are easily contaminated by multipath interferences and non-line-of-sight (NLOS) reception, leading to large positioning errors in the range of tens to hundreds meters, which is called the street canyon problem. The mainstream approaches in street canyon localization focus on extra sensors, multipath detection, and 3D maps, which still suffer from large positioning errors as well as limited application scenarios. This proposed project solves the street canyon problem with nodes’ low-level cooperation, and provides high accuracy and high availability urban positioning services. The subject of research include: (1) a satellite raw-measurement based cooperative localization approach for multi-nodes’ asynchronous satellite positioning; (2) crowdsoucing-based network topology construction; (3) performance evaluation with different network sizes, network topology changes, and external noise interferences; (4) system design and evaluation under typical application scenarios. Through theoretical analysis and real system verification, this project aims to provide a strong technical support for the street canyon problem, in the comprehensive performance of both high accuracy and high availability.