Collaborative Perception in Autonomous Driving Survey

This repo is constructed for collecting and categorizing papers about collaborative perception according to our ITSM survey paper: Collaborative Perception in Autonomous Driving: Methods, Datasets and Challenges [arXiv] [ITSM] [Zhihu]

Methods

Methods for Ideal Scenarios

• Raw data fusion
• Customized communication mechanism
• Feature fusion
• Customized loss function
• Output fusion

👉 View details in Methods for Ideal Scenarios

Methods for Real-world Issues

• Localization errors
• Communication issues
• Model or task discrepancies
• Privacy and security issues

👉 View details in Methods for Real-World Issues

Datasets

• Real-world or Simulator
• V2V or V2I

👉 View details in Datasets Summary

Challenges

• Transmission Efficiency in Collaborative Perception
• Collaborative Perception in Complex Scenes
• Federated Learning-based Collaborative Perception
• Collaborative Perception with Low Labeling Dependence

👉 View details in New Trends

Citation

If you find this work useful, please cite our paper:

@article{han2023collaborative,
  author={Han, Yushan and Zhang, Hui and Li, Huifang and Jin, Yi and Lang, Congyan and Li, Yidong},
  journal={IEEE Intelligent Transportation Systems Magazine}, 
  title={Collaborative Perception in Autonomous Driving: Methods, Datasets, and Challenges}, 
  year={2023},
  volume={15},
  number={6},
  pages={131-151},
  doi={10.1109/MITS.2023.3298534}}