Optimizing location positioning using hybrid TOA-AOA techniques in mobile cellular networks

Nikos Deligiannis, Spiros Louvros, Stavros Kotsopoulos

Abstract

Accurate mobile positioning is critical for emergency services (E-911) and location-based applications. However, standard Time of Arrival (TOA) and Angle of Arrival (AOA) techniques suffer significant accuracy degradation in urban environments due to non-Line-of-Sight (nLOS) propagation. This paper presents "Enhanced" versions of these algorithms (E-TOA and E-AOA) that integrate Geographical Information Systems (GIS) data through a Sentinel Function. By intelligently weighting or excluding nLOS signals, these hybrid methods reduce location errors by up to 40% in dense urban settings without requiring handset hardware upgrades.

Introduction

In ideal conditions, mobile devices are located via Line-of-Sight (LOS) signals. TOA measures signal travel time to determine distance (trilateration), while AOA measures the direction of incoming signals (triangulation). In real-world urban environments, signals frequently reflect off buildings and terrain, creating nLOS paths. These reflections cause the signal to travel a longer distance and arrive from an incorrect angle, leading to substantial positioning errors. While Assisted GPS (A-GPS) offers high accuracy, its cost and hardware requirements make network-based algorithmic improvements a more scalable alternative.

Methodology: The Sentinel Function and Enhanced Algorithms

The core innovation of this research is the Sentinel Function (SF). By utilizing GIS data, the network can calculate the distance to the nearest physical obstacle in any given direction from a Base Station (BS). By comparing the measured TOA distance to the distance of the nearest obstacle in the measured AOA direction, the system can identify whether a signal is likely LOS or nLOS.

E-TOA and E-AOA Implementation

  1. E-TOA (Enhanced Time of Arrival): This algorithm applies weight coefficients ($\ell_i$) to the cost function used for position estimation. Signals identified as nLOS are assigned drastically lower weights (e.g., 1/100 or 1/1000) based on the degree of deviation from the expected path.
  2. E-AOA (Enhanced Angle of Arrival): This method utilizes binary weighting. If a base station is determined to be in nLOS and likely to provide a highly erroneous angle, its contribution is discarded entirely, relying instead on more reliable LOS base stations.
Triangulation using Angle of Arrival (AOA) where BS3 is identified as a non-Line-of-Sight (nLOS) source.

Simulation Environments

To validate these methods, the authors developed three 2-D microcellular simulation environments:

  • Rural: High-speed roads with minimal building interference.
  • Suburbs: Moderate building density with antennas at lamp-post height.
  • City Centre: High building density where subscribers are located in open street grids between structures.

Each environment covers a 440x440m area, allowing for high-resolution testing of signal propagation and reflection models.

Results and Discussion

The hybrid techniques demonstrated a significant reduction in location error across all scenarios. In the dense City Centre environment, the error reduction reached 40% compared to conventional methods. Specifically, E-AOA brought the mean error down to 32.78m, a vast improvement over the standard AOA mean error of 79.02m. In rural settings, E-TOA achieved a mean error of just 7.93m.

Comparative performance of standard and enhanced positioning algorithms against true mobile locations.

Crucially, the simulation results met the U.S. Federal Communication Committee (FCC) E-911 standards, which require accuracy within 100m in 67% of cases and 300m in 95% of cases. The results indicate that AOA-based methods actually perform better in dense urban microcells than in rural areas because reflections often occur closer to the source, resulting in less angular deviation.

Conclusion

By shifting the "intelligence" of positioning to the software level at the base station and leveraging existing GIS data, the proposed E-TOA and E-AOA methods provide a cost-effective path to high-accuracy mobile location. Future work will involve expanding these models to 3-D GIS data and accounting for dynamic interference from moving objects like vehicles and foliage.

Summary updated

BibTeX
@inproceedings{deligiannis2007optimizing,
  author = {Nikos Deligiannis and Spiros Louvros and Stavros Kotsopoulos},
  title = {{Optimizing location positioning using hybrid TOA-AOA techniques in mobile cellular networks}},
  booktitle = {MobiMedia},
  year = {2007},
  doi = {10.4108/icst.mobimedia2007.1906},
}
Figure