An Integrated Positioning Method Based on Weighted Factor Graph Optimization

The intelligent transportation system requires positioning services to ensure safe driving applications. The fusion positioning method of GNSS and inertial navigation system (INS) provides high-precision location services in environments with good GNSS signal reception. However, in satellite signal denial environments such as tunnels, GNSS signals become weak or interrupted due to obstructions, resulting in a decline in observation quality or even making the signals unavailable, which leads to a sharp deterioration in positioning accuracy. Ground-based navigation system can serve as a supplement and enhancement to GNSS, offering a solution for continuous high-precision positioning both inside and outside tunnels, as well as in enclosed spaces. To address the above issues, this study proposes a weighted factor graph optimization algorithm that integrates inertial navigation and ground-based navigation. Based on the traditional factor graph model, weighted factors are introduced to reduce the impact of outlier observations on positioning results, enhance the model's anti-interference and fault tolerance capabilities, and improve the system's positioning performance. The sliding window optimization method is employed to reduce computational complexity while retaining effective historical data, improving computational speed, and ensuring real-time positioning output. Simulation results indicate that, compared with unweighted factor graph positioning schemes, the proposed integrated positioning algorithm improves positioning accuracy by 75.8%, while significantly reducing the computational load of information fusion.