A Timing Estimation Method Based-on Skewness Analysis in Vehicular Wireless Networks

Xuerong Cui, Juan Li, Chunlei Wu, Jianhang LiuSensors 2015 · DOI

TL;DR

This paper introduces a novel timing estimation method for vehicular wireless networks using the IEEE 802.11p standard. By combining cross-correlation, summing-up, and a dynamic threshold based on skewness analysis, the proposed method achieves higher precision in estimating time delays, especially in challenging low signal-to-noise ratio (SNR) and multipath environments, crucial for cooperative collision avoidance systems.

What Problem Does This Paper Address?

Accurate vehicle positioning is vital for cooperative collision avoidance (CCA) systems, aiming to reduce traffic accidents. While Global Navigation Satellite Systems (GNSS) are common, they suffer from poor reliability in urban areas or when signals are blocked (non-line-of-sight, NLOS). Short-range wireless communication, particularly the IEEE 802.11p standard designed for vehicular environments, offers a promising alternative. However, accurately estimating the time delay or range between vehicles using 802.11p signals is challenging due to multipath fading, noise, and Doppler shifts. Existing timing estimation techniques often struggle with these conditions, leading to uncertainty and reduced performance, especially at low SNRs. This paper aims to bridge this gap by proposing a more robust timing estimation method tailored for these demanding vehicular scenarios.

Methods

The proposed method, termed the Cross-correlation, Summing-up, and Skewness-analysis (CSS) method, builds upon existing techniques but introduces a novel dynamic thresholding mechanism. It involves three main steps:

  1. Cross-Correlation: Similar to the MathWorks method, it computes the cross-correlation between the received signal and the IEEE 802.11p short preamble. This step generates a timing metric, M(d), which ideally should have sharp peaks at the correct time offset.
  2. Summing-Up: To enhance the robustness against noise and multipath, the correlated samples are summed up in groups. Specifically, nine peaks of the M(d) metric, spaced 16 samples apart (corresponding to the short preamble structure), are summed to create a new metric, G(t). This step aims to consolidate the signal energy and suppress spurious peaks.
  3. Skewness-Analysis and Dynamic Threshold: This is the core innovation. Instead of using a fixed threshold, the paper proposes a dynamic threshold determined by the skewness of the G(t) metric. Skewness, a measure of the asymmetry of the probability distribution, is shown to be a monotonic function of the signal-to-noise ratio (Eb/No) in vehicular channels. By analyzing the relationship between estimation error, skewness, and threshold, the authors derive a polynomial function to dynamically set the optimal threshold (ξbest) based on the calculated skewness of G(t). This dynamic thresholding is crucial for adapting to varying noise levels and improving accuracy in low SNR conditions.

Datasets

The paper utilizes simulated data based on the International Telecommunications Union (ITU) vehicular multipath channel model. Specifically, it employs the ITU Channel A, which includes parameters for low delay spread. The simulations are conducted for various Eb/No values ranging from -15 dB to 20 dB, with 2000 channel realizations for each Eb/No to ensure statistical significance. No real-world datasets are used; the evaluation is based on these simulated channel conditions.

Key Insights and Ideas

  • Skewness as a Robust Indicator: The paper highlights that skewness of the summed correlation metric (G(t)) is a more reliable indicator of signal quality and timing accuracy in vehicular environments than standard deviation, especially across a wide range of Eb/No values. This is because skewness exhibits a monotonic behavior with Eb/No, making it suitable for dynamic thresholding.
  • Dynamic Thresholding is Crucial: Fixed thresholds, common in many timing estimation methods, perform poorly in dynamic vehicular environments with fluctuating noise levels. The proposed dynamic threshold, derived from skewness, significantly improves accuracy, particularly at low SNRs where fixed thresholds fail.
  • Summing-Up Mitigates Multipath: The summing-up step effectively consolidates the signal energy from repeated preamble symbols, making the peak detection more robust to multipath effects, which can otherwise create multiple spurious peaks or distort the main peak.
  • IEEE 802.11p Suitability: The research demonstrates the potential of IEEE 802.11p for precise ranging and positioning, going beyond its primary use for communication.

Key Results

  • Superior Performance at Low SNR: The proposed CSS-dynamic method consistently outperforms other methods (SC, MathWorks, RMB, Liu) in terms of Mean Absolute Error (MAE) at low Eb/No values. For instance, at Eb/No = -10 dB, the CSS-dynamic method achieves an MAE of 66 samples, significantly better than the next best method (CSS-fixed at 101 samples).
  • High Accuracy in Multipath: Even in the presence of multipath (Eb/No > 13 dB), the CSS-dynamic method, along with CSS-fixed and MathWorks, achieves MAEs close to 0.
  • High Percentage of Zero Error: The CSS-dynamic and CSS-fixed methods achieve the highest percentage of timing estimation error equal to zero. At Eb/No = 0 dB, they achieve nearly 59% zero error, whereas other methods require significantly higher Eb/No to reach comparable performance.
  • Robustness to Multipath: Figure 17 shows that the CSS method's timing metric remains relatively stable and less affected by multipath compared to the SC-80 method, which exhibits significant distortion.

Limitations and Future Work

  • Simulation-Based: The results are based on simulations using the ITU vehicular channel model. While this model is well-established, real-world vehicular environments can be more complex and may present additional challenges not fully captured by the simulation.
  • Coarse Timing Estimation: The paper focuses on coarse time offset estimation using the short preamble. The authors mention that this is a crucial first step for subsequent fine timing estimation and frequency shift estimation.
  • Future Work: The authors plan to extend this work to include fine timing estimation and frequency shift estimation. They also aim to apply this robust coarse timing estimation method to cooperative collision avoidance applications to reduce traffic accidents. The CSS method's applicability to other OFDM systems with repeated preamble symbols, like IEEE 802.11a, is also noted as a potential area for further exploration.

Summary updated

Abstract

Vehicle positioning technology has drawn more and more attention in vehicular wireless networks to reduce transportation time and traffic accidents. Nowadays, global navigation satellite systems (GNSS) are widely used in land vehicle positioning, but most of them are lack precision and reliability in situations where their signals are blocked. Positioning systems base-on short range wireless communication are another effective way that can be used in vehicle positioning or vehicle ranging. IEEE 802.11p is a new real-time short range wireless communication standard for vehicles, so a new method is proposed to estimate the time delay or ranges between vehicles based on the IEEE 802.11p standard which includes three main steps: cross-correlation between the received signal and the short preamble, summing up the correlated results in groups, and finding the maximum peak using a dynamic threshold based on the skewness analysis. With the range between each vehicle or road-side infrastructure, the position of neighboring vehicles can be estimated correctly. Simulation results were presented in the International Telecommunications Union (ITU) vehicular multipath channel, which show that the proposed method provides better precision than some well-known timing estimation techniques, especially in low signal to noise ratio (SNR) environments.
BibTeX
@inproceedings{cui2015timing,
  author = {Xuerong Cui and Juan Li and Chunlei Wu and Jianhang Liu},
  title = {{A Timing Estimation Method Based-on Skewness Analysis in Vehicular Wireless Networks}},
  booktitle = {Sensors},
  year = {2015},
  doi = {10.3390/s151128942},
}
Figure