Multimodal-Wireless

Multimodal-Wireless

A Large-Scale Open-Source Dataset for Multimodal Sensing and Communication Research

What is Multimodal-Wireless?

Multimodel-Wireless is a large-scale open-source dataset for multimodal sensing and communication research. It is generated through an integrated and customizable data pipeline built upon the CARLA simulator and Sionna framework, and features high-resolution communication channel state information (CSI) fully synchronized with five other sensor modalities, namely LiDAR, RGB and depth camera, inertial measurement unit (IMU) and radar, all sampled at 100 Hz.

  • ✦ 161,400 frames collected across 4 virtual towns, 16 communication scenarios, and 3 weather conditions.
  • ✦ High-resolution CSI obtained from ray tracing in Sionna and sampled at 100 Hz for diverse applications, along with 100 Hz oversampled, temporal aligned sensor data which can be easily downsampled for practical usage.
  • ✦ All-weather resilience by design.
  • ✦ Full flexibility and extensibility by simply modifying a configuration file.
  • ✦ A unified resource for vehicle-to-everything (V2X) communication and perception.

Pipeline Update Preview

🚀 Accelerated Data Generation

20× faster generation · approximately 95% less total time

The architecture update reuses the static scene and updates dynamic vehicles frame by frame in the Sionna RT workflow, reducing repeated scene preparation and loading for custom dataset generation.

Speedup reported by the authors for the same workload; detailed benchmark timings and conditions are pending. Results may vary by scenario and hardware. Release details are forthcoming.

Explore the architectureProject on GitHub

Multimodal Data Involved

Channel
Channel
LiDAR
LiDAR
Depth Map
Depth Map
GNSS
GNSS
RGB
RGB
Radar
Radar
Bounding Box
Bounding Box
IMU
IMU

License and Citation

This dataset is licensed under the Creative Commons Attribution 4.0 International License.

If the Multimodal-Wireless Dataset is helpful for your research, please kindly cite the ICC 2026 paper：

@inproceedings{mao2026mmw,
title = {Multimodal-Wireless: A Large-Scale Dataset for Sensing and Communication},
author = {Tianhao Mao and Le Liang and Jie Yang and Hao Ye and Shi Jin and Geoffrey Ye Li},
booktitle = {IEEE International Conference on Communications (ICC)},
year = {2026}
}