TL;DR
This paper introduces a novel method for improving the accuracy and speed of precise satellite positioning (ambiguity resolution) over long distances. By developing a new ionospheric model and integrating it into the existing triple-frequency carrier ambiguity resolution (TCAR) framework, the authors significantly reduce the time needed to achieve a fixed, precise solution and enhance the precision of ionospheric delay estimations.
What Problem Does This Paper Address?
Precise positioning using Global Navigation Satellite Systems (GNSS) like GPS relies on accurately determining integer ambiguities in carrier phase measurements. For long baselines (hundreds of kilometers), this process, known as Ambiguity Resolution (AR), is severely hampered by the Earth's ionosphere, which bends and delays satellite signals. Traditional methods struggle with the residual ionospheric effects, requiring large amounts of data and long processing times. While triple-frequency signals (like those from newer GPS, Galileo, and BeiDou systems) offer potential improvements, effectively mitigating the ionospheric impact remains a challenge, especially for real-time applications. This paper aims to overcome these limitations by developing a more accurate ionospheric model and integrating it as a constraint within the AR process.
Methods
The authors propose a three-pronged approach to enhance TCAR for long baselines:
- New Ionospheric Model: They introduce a novel model to estimate double-differenced (DD) ionospheric delays. This model leverages both pseudorange and phase observations from three frequencies, increasing the degrees of freedom and minimizing noise constraints for a more globally optimized ionospheric estimation.
- Smoothed Ionospheric Delay Estimates: Epoch-differenced ionospheric information is used to smooth the estimates from the new model via the Hatch algorithm. The paper also defines an optimal smoothing length (found to be 16 epochs) and proposes a strategy to diagnose and correct wrongly determined wide-lane ambiguities, which can otherwise degrade ionospheric estimates.
- Ionosphere-Weighted Model: The precisely estimated and smoothed ionospheric delays are then incorporated as pseudorange observables into the geometry-based model. This creates an "ionosphere-weighted model" that strengthens the observation model and improves the convergence efficiency of unknown parameters, particularly the ionospheric unknown.
Datasets
The experiments were conducted using real GPS data collected from six GNSS stations belonging to the MGEX network, forming three independent long baselines (HOFN-MYVA, OBE4-WTZ3, and LLAG-MASL) with lengths ranging from 104 km to 173 km. Data was collected over a 10-day period in January 2015. For cases where only two frequency signals were available, a semi-generated method was used to simulate the third frequency signal.
Key Insights and Ideas
- Ionosphere is the Bottleneck: The paper strongly emphasizes that the ionosphere is the primary obstacle to fast and reliable ambiguity resolution on long baselines.
- Triple-Frequency Synergy: Combining signals from three frequencies is crucial, but its full potential is unlocked only when the ionospheric effects are accurately modeled and constrained.
- Smoothing is Key for Precision: Smoothing ionospheric estimates over multiple epochs is essential to achieve centimeter-level precision, which is vital for effective constraints.
- Optimal Smoothing Length: There's an optimal window for smoothing ionospheric estimates; too short, and noise remains; too long, and the dynamic nature of the ionosphere is missed. The paper identifies 16 epochs as optimal.
- Constraint Power: Integrating precise ionospheric information as a constraint significantly boosts the performance of the AR process, leading to faster convergence and more reliable fixes.
Key Results
- 20% Higher Ionospheric Precision: The new ionospheric model achieves a 20% higher precision in ionospheric delay estimation compared to the current ionospheric combination method.
- Centimeter-Level Smoothing: Smoothed ionospheric estimates can achieve magnitudes defined to the nearest centimeter, enabling effective constraints.
- 20% Faster First-Fixed Solution (TFFS): The proposed ionosphere-weighted model reduces the time to achieve the first-fixed solution (TFFS) by approximately 20% compared to the geometry-based model. For example, on the HOFN-MYVA baseline, the average TFFS decreased from 43.76 epochs to 34.34 epochs.
- Improved Ambiguity Resolution: The new method leads to a higher percentage of epochs with a reciprocal ratio less than 0.33 (85.67% for the new method vs. 81.37% for the old method), indicating more reliable ambiguity fixing.
- Performance Dependence on Baseline Length: The improvement in TFFS is most pronounced for baselines around 166 km, with benefits decreasing for longer baselines (>200 km) and being less significant for shorter baselines (<100 km) where double-differencing already mitigates many errors.
Limitations and Future Work
- Model Errors: While the new method significantly improves ionospheric estimation, residual errors (potentially due to unmodeled effects) can still cause biases of up to 3 cm, though these are considered negligible for AR. The paper suggests these errors are systematic and may stem from uncorrected model errors.
- Baseline Length Impact: The effectiveness of the proposed method varies with baseline length. For very long baselines (>200 km), unmodeled errors become more dominant, and the benefit of the ionospheric constraint diminishes.
- Future Work: The authors suggest that their method is particularly beneficial for systems like BeiDou, which offer triple-frequency signals on all satellites. The real-time estimation of ionospheric delay also opens possibilities for real-time relative positioning and direct ionospheric studies. The paper also implies that further refinement of the ionospheric model itself could lead to even greater improvements.
Summary updated