Welcome to the Syndra main page

SYNDRA is a train environments for realistic simulation.

We have implemented SYNDRA to generate labelled visual datasets (lidar and camera) for testing perceptual algorithms in railway environments.

The railway sector has historically been fundamental for global transportation
Achieving advancements in railway functions requires enhancing certain tasks and automating others:

- Train odometry
- Localization
- Track discrimination
- Obstacle detection
- Segmentation

Simultaneously, Technological Advancements:

- Rapid growth in automotive and robotics leads to significant improvements in perceptual-based understanding using visual sensors.
- Perceptual algorithms are essential for object detection, scene understanding, and navigation.
- Development relies on large, diverse, and accurately annotated datasets especially for AI algorithms.

A notable gap exists between the robotics and automotive sectors and the railway domain regarding these technologies.
Specifically, the railway sector is characterized by a lack of open (real or synthetic) datasets that slows down the innovation and development in this area.

To fill this gap a simulation frameworks based on Unreal Engine 5 have been implemented:

- Generate different scenarios
- Generate train routes
- Generate objects like tunnels, bridges, and vegetation
- Acquire data from visual sensors as cameras and LiDARS
- Automatic data labeling for different tasks

The simulation framework is not open-sourced yet, but here we provide datasets for different tasks.

Semantic Segmentation

Different adversarial condition and seasonal changes

If you are interested in synthetic data generation for railway environments, we are always open to new collaboration, check our previous work:

1. D’Amico et al. “A Comparative Analysis of Visual Odometry in Virtual and Real-World Railways Environments.”
   Accepted at RAILWAYS 2024: The Sixth International Conference on Railways Technology, September 2024, Prague
      
2. D’Amico et al. "TrainSim: A railway simulation framework for LiDAR and camera dataset generation." IEEE Transactions on Intelligent Transportation Systems (2023)
      
3. D'Amico et al. “Graphic Simulation Framework of Railway Scenarios for LiDAR Dataset Generation
   RAILWAYS 2022: The Fifth International Conference on Railways Technology, August 2022, Montpellier