Intelligent 3D mapping aided GNSS based collaborative positioning in urban areas

Abstract

Localization is essential for almost every civil application, mainly relying on the global navigation satellite system (GNSS). Consumer-grade GNSS receivers experience different types of interference and noise, resulting in unsatisfactory positioning accuracy. Owing to the development of communication technologies, the concept of collaborative positioning can be applied to GNSS to effectively reduce positioning errors. However, this algorithm is ineffective for urban areas where most applications are located, owing to the severe degradation from multipath and non-line-of-sight (NLOS) reception errors. This study develops a novel 3D mapping-aided (3DMA) GNSS collaborative positioning algorithm effective in urban areas. It complementarily integrates the 3DMA GNSS algorithm and the double difference (DD) based estimation to eliminate common errors and mitigate distinctive errors simultaneously, thereby providing better relative position information to optimize the urban GNSS solutions. In this thesis, the proposed 3DMA GNSS collaborative positioning algorithm is developed and analyzed comprehensively. The performance and limitations of conventional collaborative positioning algorithms in urban areas were first evaluated. Then, a preliminary 3DMA GNSS collaborative positioning algorithm with multipath and NLOS exclusion was developed. Next, an improved 3DMA GNSS collaborative positioning algorithm, utilizing NLOS receptions as features to aid positioning, was developed to be effective in dense urban areas. Subsequently, the practical issues of the proposed algorithm were analyzed through realistic simulations, including the scalability performance and latency degradation. Finally, two collaborator selection strategies for improving the algorithm effectiveness are investigated based on the environmental context and occurrence of spatial correlation. Extensive simulations and experiments were conducted to validate the performance of the proposed algorithm, which outperformed the conventional positioning methods with approximately twice the accuracy. Therefore, the proposed 3DMA GNSS collaborative positioning algorithm is capable of providing accurate and robust positioning solutions for agents in dense urban areas.