Scholarlink

<p>In disrupted satellite environment, traditional filtering methods struggle to maintain stable positioning solutions by effectively utilizing historical information. This study proposes a positioning method based on factor graph optimization (FGO), which constructs factor graph nodes using weighted pseudorange measurements and the Doppler single-point velocity results from BeiDou and GPS satellite systems, to achieve satellite positioning outcomes. Further, these positioning results can be integrated with the pre-integration results from micro-electro-mechanical system inertial measurement unit to form new factor graph nodes, enhancing the positioning process. Experimental results indicate that the proposed FGO method achieves higher positioning accuracy in both global navigation satellite system (GNSS) single point positioning (SPP) and loosely coupled scenarios compared to traditional methods. In scenarios with an artificially induced 30-second satellite signal interruption, the FGO loosely coupled method shows a 0.2% reduction in horizontal positioning error and a 32% reduction in vertical positioning error compared to the GNSS SPP method, while the traditional filtering loosely coupled method exhibits error increases of 8.4% and 55%, respectively.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Quantitative Evaluation of Hybrid GNSS-Only and GNSS-IMU Factor Graph Optimization

  • 期刊论文

  • 发表日期：

    2025年12月1日

查看全文
  • Yiqian Li,
  • Di He

摘要

In disrupted satellite environment, traditional filtering methods struggle to maintain stable positioning solutions by effectively utilizing historical information. This study proposes a positioning method based on factor graph optimization (FGO), which constructs factor graph nodes using weighted pseudorange measurements and the Doppler single-point velocity results from BeiDou and GPS satellite systems, to achieve satellite positioning outcomes. Further, these positioning results can be integrated with the pre-integration results from micro-electro-mechanical system inertial measurement unit to form new factor graph nodes, enhancing the positioning process. Experimental results indicate that the proposed FGO method achieves higher positioning accuracy in both global navigation satellite system (GNSS) single point positioning (SPP) and loosely coupled scenarios compared to traditional methods. In scenarios with an artificially induced 30-second satellite signal interruption, the FGO loosely coupled method shows a 0.2% reduction in horizontal positioning error and a 32% reduction in vertical positioning error compared to the GNSS SPP method, while the traditional filtering loosely coupled method exhibits error increases of 8.4% and 55%, respectively.