Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Computer Vision Approach for Autonomous Cars to Drive Safe at Construction Zone

Abu Shad Ahammed, Md Shahi Amran Hossain, Roman Obermaisser

TL;DR

This paper tackles safe autonomous driving in construction zones by addressing drift conditions that degrade perception. It presents a drift-aware obstacle-detection pipeline based on YOLOv8 trained on CARLA-generated data for cones, beacons, and barriers, with data augmentation to simulate adverse conditions and achieving a validation mAP around 0.97 and sub-2 ms inference. The authors provide a detailed data-generation and labeling workflow, plus strong per-class performance on both validation and test sets, suggesting readiness for real-time ADS integration. Overall, the work advances safe navigation in dynamic work zones by demonstrating robust, drift-aware CV methods suitable for safety-critical automotive applications.

Abstract

To build a smarter and safer city, a secure, efficient, and sustainable transportation system is a key requirement. The autonomous driving system (ADS) plays an important role in the development of smart transportation and is considered one of the major challenges facing the automotive sector in recent decades. A car equipped with an autonomous driving system (ADS) comes with various cutting-edge functionalities such as adaptive cruise control, collision alerts, automated parking, and more. A primary area of research within ADAS involves identifying road obstacles in construction zones regardless of the driving environment. This paper presents an innovative and highly accurate road obstacle detection model utilizing computer vision technology that can be activated in construction zones and functions under diverse drift conditions, ultimately contributing to build a safer road transportation system. The model developed with the YOLO framework achieved a mean average precision exceeding 94\% and demonstrated an inference time of 1.6 milliseconds on the validation dataset, underscoring the robustness of the methodology applied to mitigate hazards and risks for autonomous vehicles.

A Computer Vision Approach for Autonomous Cars to Drive Safe at Construction Zone

TL;DR

This paper tackles safe autonomous driving in construction zones by addressing drift conditions that degrade perception. It presents a drift-aware obstacle-detection pipeline based on YOLOv8 trained on CARLA-generated data for cones, beacons, and barriers, with data augmentation to simulate adverse conditions and achieving a validation mAP around 0.97 and sub-2 ms inference. The authors provide a detailed data-generation and labeling workflow, plus strong per-class performance on both validation and test sets, suggesting readiness for real-time ADS integration. Overall, the work advances safe navigation in dynamic work zones by demonstrating robust, drift-aware CV methods suitable for safety-critical automotive applications.

Abstract

To build a smarter and safer city, a secure, efficient, and sustainable transportation system is a key requirement. The autonomous driving system (ADS) plays an important role in the development of smart transportation and is considered one of the major challenges facing the automotive sector in recent decades. A car equipped with an autonomous driving system (ADS) comes with various cutting-edge functionalities such as adaptive cruise control, collision alerts, automated parking, and more. A primary area of research within ADAS involves identifying road obstacles in construction zones regardless of the driving environment. This paper presents an innovative and highly accurate road obstacle detection model utilizing computer vision technology that can be activated in construction zones and functions under diverse drift conditions, ultimately contributing to build a safer road transportation system. The model developed with the YOLO framework achieved a mean average precision exceeding 94\% and demonstrated an inference time of 1.6 milliseconds on the validation dataset, underscoring the robustness of the methodology applied to mitigate hazards and risks for autonomous vehicles.
Paper Structure (12 sections, 3 figures, 3 tables)

This paper contains 12 sections, 3 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Related Research
  3. Construction Zone ODD
  4. HARA in Construction Zone
  5. Data Organization and Pre-processing
  6. Data Acquisition from CARLA
  7. Data Filtering
  8. Data Augmentation and Labeling
  9. Data Segmentation
  10. Object Detection Modeling for CZ
  11. Results and Discussion
  12. Conclusion

Figures (3)

  • Figure 1: Road Obstructions in Construction Zones
  • Figure 2: An Overview of Data Management and Pre-processing
  • Figure 3: Performance Curves of the Developed Model