PPP/INS Tight Integration with BDS-3 PPP-B2b Service in the Urban Environment

Luguang Lai, Xin Meng, Dongqing Zhao, Xin Li, Wenzhuo Guo, Linyang Li2023 · DOI

TL;DR

This paper presents a tightly integrated system combining BeiDou-3's PPP-B2b real-time correction service with an Inertial Navigation System (INS) for robust positioning in challenging urban environments. The research demonstrates that this integration achieves decimeter-level accuracy, crucial for applications like autonomous driving, even during GNSS signal interruptions, and highlights the significant impact of IMU quality on overall performance.

What Problem Does This Paper Address?

Precise and reliable real-time positioning is increasingly critical for advanced applications such as autonomous driving, robotics, and the Internet of Things. While Global Navigation Satellite Systems (GNSS) like BeiDou-3 (BDS-3) offer precise point positioning (PPP), their performance can be severely degraded in urban environments due to signal blockage, multipath effects, and internet connectivity issues for real-time correction services. Traditional PPP often relies on post-processing, introducing latency. Real-time kinematic (RTK) requires extensive infrastructure. To overcome these limitations, this paper addresses the challenge of achieving continuous and reliable decimeter-level positioning in urban settings by tightly integrating BDS-3's PPP-B2b service with an Inertial Navigation System (INS). This integration aims to leverage the strengths of both systems: GNSS for absolute positioning and INS for short-term, high-frequency motion sensing, especially during GNSS outages.

Methods

The core of this research is the development and evaluation of a tightly integrated PPP-B2b/INS model. This approach differs from loosely coupled systems by fusing GNSS and INS measurements at the state estimation level, allowing for more robust error handling and improved performance during GNSS signal interruptions.

Key methodological aspects include:

  • BDS-3 PPP-B2b Service Utilization: The paper leverages the real-time corrections broadcast by BDS-3's PPP-B2b signal. These corrections are used to mitigate satellite orbit and clock errors, enabling real-time PPP.
  • Ionosphere-Free (IF) Combination: To eliminate the first-order ionospheric delay, an ionosphere-free combination of GNSS measurements is employed.
  • INS Error Modeling: A standard strapdown INS error model is used, considering position, velocity, and misalignment errors. Biases of accelerometers and gyroscopes are modeled as random walk processes.
  • Tight Integration: The GNSS and INS state equations are fused into a single system. The GNSS observation equations are augmented with INS-predicted observations. This tight coupling allows the INS to provide continuous position estimates during GNSS outages, while GNSS data aids in correcting INS drift.
  • Quality Control: A detection, identification, and adaptation (DIA) strategy is employed for quality control of the GNSS observations.

Datasets

The study utilizes data collected from a vehicle-mounted experiment conducted in Zhengzhou, China, on October 18, 2022. The experiment lasted approximately 1.5 hours and involved various driving maneuvers (acceleration, deceleration, turning) in an urban environment.

  • GNSS Data: Raw GNSS observations were recorded at 1 Hz.
  • IMU Data: Data from two types of Inertial Measurement Units (IMUs) were recorded:
  • A tactical IMU (SNC300A-DGI) at 200 Hz.
  • A MEMS IMU (ADIS-16470) at 100 Hz.
  • Reference Data: A reference station with a tightly coupled multi-GNSS RTK/INS system using commercial software (Inertial Explorer 8.9) was set up nearby to provide a smooth solution for comparison and time alignment.

Key Insights and Ideas

  • Urban Navigation is Tough: The urban environment poses significant challenges for GNSS, with tall buildings, trees, and viaducts causing signal blockage and multipath, leading to errors of several meters or even complete positioning interruptions for standalone PPP.
  • Tight Integration is Key for Continuity: Tightly coupling PPP-B2b with INS is crucial for maintaining continuous and reliable positioning during short GNSS signal interruptions. The INS bridges the gaps left by GNSS, preventing complete loss of position.
  • IMU Quality Matters Dramatically: The performance of the integrated system is highly dependent on the quality of the IMU. A tactical-grade IMU yields significantly better results than a low-cost MEMS IMU, especially in terms of attitude determination and during GNSS outages.
  • BDS-3 PPP-B2b is Promising but Needs Growth: The BDS-3 PPP-B2b service shows promise for real-time positioning without internet reliance, but its current performance is slightly behind commercial RT products (like GFZ-SSR) and significantly behind post-processed products (like GFZ-GBM) in terms of accuracy. This is attributed to the limited number of available satellites and potentially less mature correction products.
  • Real-time vs. Post-processing: For velocity and attitude accuracy, there's no significant difference between using real-time (BDS-3 PPP-B2b, GFZ-SSR) and post-processed (GFZ-GBM) products when using a high-quality IMU. This suggests that the IMU's precision is the dominant factor for these parameters.

Key Results

  • Positioning Accuracy:
    • The tightly integrated PPP-B2b/INS achieved decimeter-level positioning accuracy in the urban environment: E, N, U components were 0.292 m, 0.115 m, and 0.155 m, respectively.
  • Compared to GFZ-SSR (0.183 m, 0.098 m, 0.122 m), the BDS-3 PPP-B2b/INS had a 3D positioning gap of about 1 dm.
  • Compared to GFZ-GBM (0.115 m, 0.055 m, 0.056 m), the gap was about 2 dm.
  • Velocimetry Accuracy:
    • With a tactical IMU, the velocimetry accuracy was consistently around 0.3 cm/s for E, N, and U components, regardless of the precise product used.
  • Attitude Accuracy:
    • With a tactical IMU, pitch and roll accuracy were better than 0.01 deg, and yaw accuracy was around 0.1 deg.
  • MEMS IMU Performance:
    • Using a MEMS IMU significantly degraded performance. Positioning accuracy worsened by 10-40%, and attitude accuracy decreased most significantly (pitch/roll around 0.6/0.3 deg, yaw around 1.1 deg).

Limitations and Future Work

  • Limited Satellite Support: The current PPP-B2b service primarily supports BDS-3 and GPS, with a limited number of available satellites (average 12.34 in this experiment) compared to commercial products. This impacts overall accuracy.
  • INS Error Accumulation: The paper acknowledges that INS error accumulates rapidly during prolonged GNSS signal interruptions, which can degrade performance.
  • Future Work Suggestions:
    • Improved PPP-B2b: With the development of BDS-3, support for more satellites and systems is expected, leading to improved PPP-B2b correction precision and thus better integrated system performance.
  • Sensor Fusion: To further mitigate INS error accumulation during long GNSS outages, integrating additional sensors like odometers and cameras is proposed.
  • Integrity Monitoring: Developing robust integrity monitoring for the PPP/INS system is identified as a crucial area for future research.

Summary updated

BibTeX
@misc{lai2023ppp,
  author = {Luguang Lai and Xin Meng and Dongqing Zhao and Xin Li and Wenzhuo Guo and Linyang Li},
  title = {{PPP/INS Tight Integration with BDS-3 PPP-B2b Service in the Urban Environment}},
  year = {2023},
  doi = {10.3390/s23052652},
}
Figure