Journal of information and communication convergence engineering

The Korea Institute of Information and Commucation Engineering (한국정보통신학회)
권호 표지
DOI QR코드

DOI QR Code

Effects of CNN Backbone on Trajectory Prediction Models for Autonomous Vehicle

  • Seoyoung Lee (Department of Computer Science and Engineering, Korea University of Technology and Education) ;
  • Hyogyeong Park (Department of Computer Science and Engineering, Korea University of Technology and Education) ;
  • Yeonhwi You (Department of Computer Science and Engineering, Korea University of Technology and Education) ;
  • Sungjung Yong (Department of Computer Science and Engineering, Korea University of Technology and Education) ;
  • Il-Young Moon (Department of Computer Science and Engineering, Korea University of Technology and Education)
  • Received : 2023.07.08
  • Accepted : 2023.10.23
  • Published : 2023.12.31
Download PDF
⟨ Previous Next ⟩

Abstract

Trajectory prediction is an essential element for driving autonomous vehicles, and various trajectory prediction models have emerged with the development of deep learning technology. Convolutional neural network (CNN) is the most commonly used neural network architecture for extracting the features of visual images, and the latest models exhibit high performances. This study was conducted to identify an efficient CNN backbone model among the components of deep learning models for trajectory prediction. We changed the existing CNN backbone network of multiple-trajectory prediction models used as feature extractors to various state-of-the-art CNN models. The experiment was conducted using nuScenes, which is a dataset used for the development of autonomous vehicles. The results of each model were compared using frequently used evaluation metrics for trajectory prediction. Analyzing the impact of the backbone can improve the performance of the trajectory prediction task. Investigating the influence of the backbone on multiple deep learning models can be a future challenge.

Keywords

Acknowledgement

This research was supported by the Basic Research Program through the National Research Foundation of Korea (NRF), funded by the Ministry of Education (No. 2021R1I1A3057 800), and the results were supported by the Regional Innovation Strategy (RIS) through the NRF funded by the Ministry of Education (MOE) (2021RIS-004).

References

  1. B. D. Kim, S. H. Park, S. H. Lee, E. Khoshimjonov, D. S. Kum, J. S. Kim, J. S. Kim, and J. W. Choi, "LaPred: Lane-Aware Prediction of Multi-Modal Future Trajectories of Dynamic Agents," in 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Nashville: TN, pp. 14631-14640, 2021. DOI: 10.1109/CVPR46437.2021.01440.
  2. Y. Chai, B. Sapp, M. Bansal, and D. Anguelov, "MultiPath: Multiple Probabilistic Anchor Trajectory Hypotheses for Behavior Prediction," Proceedings of the Conference on Robot Learning, vol. 100, pp. 86-99, 2020. DOI: 10.48550/arXiv.1910.05449.
  3. T. Phan-Minh, E. C. Grigore, F. A. Boulton, O. Beijbom, and E. M. Wolff, "CoverNet: Multimodal Behavior Prediction Using Trajectory Sets," in 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Seattle: WA, pp. 14062-14071, 2020. DOI: 10.1109/CVPR42600.2020.01408.
  4. H. Cui, Vl. Radosavljevic, F-C. Chou, T-H. Lin, T. Nguyen, T-K. Huang, J. Schneider, and N. Djuric, "Multimodal Trajectory Predictions for Autonomous Driving using Deep Convolutional Networks," in 2019 International Conference on Robotics and Automation (ICRA), Montreal: QC, pp. 2090-2096, 2019. DOI: 10.1109/ICRA.2019.8793868.
  5. S. Mozaffari, O. Y. Al-Jarrah, M. Dianati, P. Jennings, and A. Mouzakitis, "Deep Learning-Based Vehicle Behavior Prediction for Autonomous Driving Applications: A Review," IEEE Transactions on Intelligent Transportation Systems, vol. 23, no. 1, pp. 33-47, Jan. 2022. DOI: 10.1109/TITS.2020.3012034.
  6. K. He, X. Zhang, S. Ren, and J. Sun, "Deep residual learning for image recognition," arXiv preprint arXiv:1512.03385, Dec. 2015. DOI: 10.48550/arXiv.1512.03385.
  7. S. Xie, R. Girshick, P. Dollar, Z. Tu, and K. He, "Aggregated Residual Transformations for Deep Neural Networks," arXiv preprint arXiv:1611.05431, Nov. 2016. DOI: 10.48550/arXiv.1611.05431.
  8. S. Zagoruyko and N. Komodakis, "Wide Residual Networks," arXiv preprint arXiv:1605.07146, May 2016. DOI: 10.48550/arXiv.1605.07146.
  9. K. He, X. Zhang, S. Ren, and J. Sun, "Deep residual learning for image recognition," arXiv preprint arXiv:1512.03385, Dec. 2015. DOI: 10.48550/arXiv.1512.03385.
  10. M. Sandler, A. Howard, M. Zhu, A. Zhmognimov, and L-C. Chen, "MobileNetV2: Inverted Residual and Linear Bottlenecks," arXiv preprint arXiv:1801.04381, Jan. 2018. DOI: 10.48550/arXiv.1801.04381.
  11. nuScenes Contributors, nuScenes [Internet], Available: https://www.nuscenes.org/.
  12. D. P. Kingma, J. Ba, "Adam: A Method for Stochastic Optimization," arXiv preprint arXiv:1412.6980, Dec. 2014. DOI: 10.48550/arXiv.1412.6980.
  13. H. Caesar, V. Bankiti, A. H. Lang, S. Vora, V. E. Liong, Q. Xu, A. Krishnan, Y. Pan, G. Baldan, and O. Beijbom, "nuScenes: A multimodal dataset for autonomous driving," arXiv preprint arXiv: 1903.11027, Mar. 2019. DOI: 10.48550/arXiv.1903.11027.