Autonomous vehicles are crucial for improving road safety by reducing human errors that cause accidents. This paper presents the development of an autonomous vehicle system for obstacle avoidance and traffic sign recognition, using the NVIDIA Jetson Nano platform. The system leverages the YOLO (You Only Look Once) object detection model for real-time traffic sign recognition and obstacle avoidance. The YOLO model is trained on multiple datasets of real-world scenarios, enabling it to detect various objects such as cars, pedestrians, and obstacles that includes various lighting conditions and environments. When an obstacle is detected, the vehicle's control system is adjusted to perform appropriate actions, ensuring collision avoidance. In traffic sign recognition, YOLO can be used, and it is designed and trained using TensorFlow. TensorFlow is a popular deep learning framework used for building and training YOLO models, making it an effective choice for traffic sign recognition tasks. The NVIDIA Jetson Nano's powerful GPU makes the YOLO object detection and traffic sign recognition models work much faster. This allows the autonomous vehicle to process information in real-time and make quick decisions.
A Deep Transfer Learning Approach for Enhanced Traffic Sign Recognition in Autonomous Vehicles with NVIDIA Jetson Nano
2023-09-14
1157204 byte
Conference paper
Electronic Resource
English
Edge Computing-Driven Real-Time Drone Detection Using YOLOv9 and NVIDIA Jetson Nano
DOAJ | 2024
|