DeepSense-6G-V2I-CSI-Compression

This repository contains the code implementation for the paper "Deep Learning for Multi-Modal Sensor Fusion and CSI Compression in Vehicular Communications" by Shubham Srivastava, Marian Temprana Alonso, Rounak Chatterjee, Nurassyl Askar, Umut Demirhan, Farhad Shirani, Stefano Rini, and Ahmed Alkhateeb.

Requirements

• Python 3.x
• PyTorch
• Additional dependencies (see requirements.txt)

Installation

1. Clone the repository:

   git clone https://github.com/shubhamsrivast4u/DeepSense-6G-V2I-CSI-Compression.git
   cd DeepSense-6G-V2I-CSI-Compression
   

2. Install dependencies:

   pip install -r requirements.txt
   

   • consider also installing torch separately

Dataset

Download the DeepSense 6G dataset from DeepSense 6G Scenario 1 and place it in the root directory.

Pre-trained Models

Pre-trained model weights are available here. Download and place them in the models directory.

Usage

Our repository includes various evaluation scripts for different experimental setups:

• Experiment2_<ModelName>_without_MSI-r<X>.ipynb: Evaluates model performance without Multi-Sensor Information (MSI) for reduction rate X in Experiment 2.
• Experiment2_<ModelName>-MSI-r<X>.ipynb: Evaluates model performance with MSI for reduction rate X in Experiment 2.
• Experiment3_<ModelName>_without_MSI-r<X>.ipynb: Evaluates model performance without MSI for reduction rate X in Experiment 3.
• Experiment3_<ModelName>-withMSI-r<X>.ipynb: Evaluates model performance with MSI for reduction rate X in Experiment 3.

Replace <ModelName> with the specific model and <X> with the reduction rate.

Contact

For inquiries, please contact the corresponding author:

Shubham Srivastava - [email protected]