A Cluster-Based Weighted Feature Similarity Moving Target Tracking Algorithm for Automotive FMCW Radar

TL;DR

A moving-target tracking algorithm for automotive FMCW mmWave radar: sparse radar points are merged into clusters and matched across frames by weighted feature similarity, staying robust under strong environmental noise and multiple interfering targets.

Key contributions:

• A cluster-based weighted feature similarity matching algorithm that raises the same-target matching rate across adjacent frames.
• Trajectory extraction and correction for moving targets using the ego vehicle’s motion parameters.
• Verified in autonomous-driving experiments with high recognition accuracy and low positional error.

Learn more

Implementation details are documented on the project page.

BibTeX

@inproceedings{chen2022cluster,
  title={A Cluster-Based Weighted Feature Similarity Moving Target Tracking Algorithm for Automotive FMCW Radar},
  author={Chen, Rongqian and Zou, Yingquan and Gao, Anyong and Chen, Leshi},
  booktitle={2022 IEEE 95th Vehicular Technology Conference:(VTC2022-Spring)},
  pages={1--5},
  year={2022},
  organization={IEEE}
}