Real-Time Visual Perception Acceleration for Autonomous Driving Based on TensorRT Quantization and Mixed-Precision Inference

Tianlang Xin

doi:10.47297/taposatWSP2633-456934.20260708

PDF
Export
Share
Collection
Album

Real-Time Visual Perception Acceleration for Autonomous Driving Based on TensorRT Quantization and Mixed-Precision Inference

Volume 7, Issue 8

Views: 0 Downloads: 0

Real-Time Visual Perception Acceleration for Autonomous Driving Based on TensorRT Quantization and Mixed-Precision Inference
Real-Time Visual Perception Acceleration for Autonomous Driving Based on TensorRT Quantization and Mixed-Precision Inference
Electrical & Electronic Engineering, Xiamen University Malaysia, Selangor, 43900, Malaysia
Vol. 7, Issue 8, Pages: 172-178(2026)
Published：30 August 2026
DOI：10.47297/taposatWSP2633-456934.20260708
Accepted：

Scan QR Code

Xin Tianlang.Real-Time Visual Perception Acceleration for Autonomous Driving Based on TensorRT Quantization and Mixed-Precision Inference[J].Theory and Practice of Science and Technology,2026,07(08):172-178.
DOI：

Xin Tianlang.Real-Time Visual Perception Acceleration for Autonomous Driving Based on TensorRT Quantization and Mixed-Precision Inference[J].Theory and Practice of Science and Technology,2026,07(08):172-178. DOI： 10.47297/taposatWSP2633-456934.20260708.

摘要

Abstract

Autonomous driving systems require efficient and reliable visual perception models for real-time deployment on resource-constrained platforms. This study proposes a TensorRT-based acceleration framework for YOLOv3 by integrating mixed-precision inference and INT8 quantization to improve deployment efficiency while preserving detection performance. Experiments on the KITTI dataset show that TensorRT optimization can significantly reduce inference cost and resource consumption. FP16 inference maintains stable detection accuracy while improving efficiency

whereas INT8 quantization provides greater acceleration with only limited performance degradation. Overall

the proposed framework achieves a practical balance between accuracy and real-time performance

providing an effective solution for deploying deep learning-based visual perception models in autonomous driving scenarios.

关键词

Keywords

references

Zhao Jingyuan, Wu Yuyan, Deng Rui, et al. A Survey of Autonomous Driving from a Deep Learning Perspective[J]. ACM Computing Surveys, 2025, 57(10): 1-60.

Abrecht S, Hirsch A, Raafatnia S, et al. Deep Learning Safety Concerns in Automated Driving Perception[J]. IEEE Transactions on Intelligent Vehicles, 2025, 10(2): 1455-1466.

Boumendil A, Bechkit W, Benatchba K. On-Device Deep Learning: Survey on Techniques Improving Energy Efficiency of DNNs[J]. IEEE Transactions on Neural Networks and Learning Systems, 2025, 36(5): 7806-7821.

Li Z, Xiao J, Yang L, et al. RepQ-ViT: Scale Reparameterization for Post-Training Quantization of Vision Transformers[C]//2023 IEEE/CVF International Conference on Computer Vision (ICCV). Paris: IEEE, 2023: 17181-17190.

arXiv:1804.02767 [cs.CV][EB/OL]. https://arxiv.org/abs/1804.02767.

Hnewa M, Radha H. Object Detection Under Rainy Conditions for Autonomous Vehicles: A Review of State-of-the-Art and Emerging Techniques[J]. IEEE Signal Processing Magazine, 2021, 38(1): 53-67.

Li Z, Li H, Meng L. Model Compression for Deep Neural Networks: A Survey[J]. Computers, 2023, 12(3): 60.

Jacob B, Kligys S, Chen B, et al. Quantization and Training of Neural Networks for Efficient Integer-Arithmetic-Only Inference[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2018: 2704-2713.

Jeong E, Kim J, Ha S. TensorRT-Based Framework and Optimization Methodology for Deep Learning Inference on Jetson Boards[J]. ACM Transactions on Embedded Computing Systems, 2022, 21(5): 1-26.

Feng D, Harakeh A, Waslander S L, et al. A Review and Comparative Study on Probabilistic Object Detection in Autonomous Driving[J]. IEEE Transactions on Intelligent Transportation Systems, 2022, 23(8): 9961-9980.

Behley J, Garbade M, Milioto A, et al. SemanticKITTI: A Dataset for Semantic Scene Understanding of LiDAR Sequences[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision. 2019: 9297-9307.

Geiger A, Lenz P, Urtasun R. Are We Ready for Autonomous Driving? The KITTI Vision Benchmark Suite[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR). 2012: 3354-3361.

Garikapati D, Shetiya S S. Autonomous Vehicles: Evolution of Artificial Intelligence and the Current Industry Landscape[J]. Big Data and Cognitive Computing, 2024, 8(4): 42.

Gai W, Liu Y, Zhang J, et al. An Improved Tiny YOLOv3 for Real-Time Object Detection[J]. Systems Science & Control Engineering, 2021, 9(1): 314-321.

Alert me when the article has been cited

Submit

A Comparative Review of Artificial Intelligence for the Control of Physical Systems: Case Studies in Power-Grid Dispatch and Autonomous Driving

A Review of the Research on Pedestrian Detection of Autonomous Driving Under Complex Lighting Conditions

Multi-Modal Fusion for 3D Object Detection in Autonomous Driving: A Review

A Review of Deep Learning-Based Industrial Robot Grasping Systems

Research on Perception and Decision-Making Methods for Autonomous Driving

Related Author

Guohao Yin Hefei No.8 High School

Hanmou Sun Chongqing University of Posts and Telecommunications

Haozhou Fei Nanjing University of Posts and Telecommunications

Ziyang Yan School of Mechanical Engineering, Hefei University of Technology,Mechanical Design, Manufacturing, and Automation

Qinwen Qiu School of Management and engineering, Nanjing University, Industrial Intelligence and System Integration

Yuexin Pan School of Sports Engineering, Beijing Sport University

Huijun Chen College of Energy and Electrical Engineering, Chang'an University

Yanliang Liu Wuhan University of Technology

Related Institution

Hefei No.8 High School

Chongqing University of Posts and Telecommunications

Nanjing University of Posts and Telecommunications

School of Mechanical Engineering, Hefei University of Technology,Mechanical Design, Manufacturing, and Automation

School of Management and engineering, Nanjing University, Industrial Intelligence and System Integration

⁰