Path-Constrained State Estimation for Rail Vehicles

Cornelius von Einem,Andrei Cramariuc,Roland Siegwart,Cesar Cadena,Florian Tschopp

2023-08-23

Abstract:Globally rising demand for transportation by rail is pushing existing infrastructure to its capacity limits, necessitating the development of accurate, robust, and high-frequency positioning systems to ensure safe and efficient train operation. As individual sensor modalities cannot satisfy the strict requirements of robustness and safety, a combination thereof is required. We propose a path-constrained sensor fusion framework to integrate various modalities while leveraging the unique characteristics of the railway network. To reflect the constrained motion of rail vehicles along their tracks, the state is modeled in 1D along the track geometry. We further leverage the limited action space of a train by employing a novel multi-hypothesis tracking to account for multiple possible trajectories a vehicle can take through the railway network. We demonstrate the reliability and accuracy of our fusion framework on multiple tram datasets recorded in the city of Zurich, utilizing Visual-Inertial Odometry for local motion estimation and a standard GNSS for global localization. We evaluate our results using ground truth localizations recorded with a RTK-GNSS, and compare our method to standard baselines. A Root Mean Square Error of 4.78 m and a track selectivity score of up to 94.9 % have been achieved.

Robotics,Signal Processing

What problem does this paper attempt to address?

The paper is primarily dedicated to addressing the issue of how to achieve high-precision, continuous position estimation for railway vehicles in situations where Global Navigation Satellite System (GNSS) signals are limited or unreliable. Specifically, the goals of the paper are:

1. **Develop an accurate, robust, and high-frequency positioning system**: To cope with the growing demand for railway transportation worldwide and the existing infrastructure nearing its capacity limits, it is necessary to develop an accurate positioning system that ensures the safe and efficient operation of trains.

2. **Fuse multiple sensor data**: Since a single sensor mode is difficult to meet robustness and safety requirements, it is necessary to combine multiple sensor data to improve the system's reliability and safety.

3. **Utilize the unique characteristics of the railway network**: The paper proposes a sensor fusion framework under path constraints, leveraging the uniqueness of the railway network (i.e., trains can only move along the tracks), modeling the state space as one-dimensional track geometry.

4. **Multi-hypothesis tracking**: Considering the various travel paths a train might take within the railway network, the paper introduces a new multi-hypothesis tracking method to account for multiple possible trajectories a train might take through the railway network.

5. **Improve positioning accuracy and reliability**: By demonstrating the reliability and accuracy of the proposed method on real-world datasets, the paper aims to improve positioning accuracy and ensure the system can correctly identify the track on which the train is located.

6. **Evaluate performance**: The paper validates the effectiveness of the proposed fusion framework through actual datasets and compares its results with standard baselines, proving the method's capability in enhancing positioning accuracy and track selectivity.

In summary, the main objective of this paper is to develop a novel state estimation framework suitable for railway vehicles to ensure accurate and continuous positioning even in poor GNSS signal conditions, which is crucial for the safe and efficient operation of modern railway transportation.