Open Access

ARTICLE

Constructing a Deep Image Analysis System Based on Self-Driving and AIoT

Wen-Tsai Sung¹, Sung-Jung Hsiao^2,*, Chung-Yen Hsiao¹

1 Department of Electrical Engineering, National Chin-Yi University of Technology, Taichung, 411030, Taiwan
2 Department of Information Technology, Takming University of Science and Technology, Taipei City, 11451, Taiwan

* Corresponding Author: Sung-Jung Hsiao. Email:

Intelligent Automation & Soft Computing 2022, 31(2), 1223-1240. https://doi.org/10.32604/iasc.2022.020746

Received 06 June 2021; Accepted 12 July 2021; Issue published 22 September 2021

Abstract

This research is based on the system architecture of Edge Computing in the AIoT (Artificial Intelligence & Internet of Things) field. In terms of receiving data, the authors proposed approach employed the camera module as the video source, the ultrasound module as the distance measurement source, and then compile C++ with Raspberry Pi 4B for image lane analysis, while Jetson Nano uses the YOLOv3 algorithm for image object detection. The analysis results of the two single-board computers are transmitted to Motoduino U1 in binary form via GPIO, which is used for data integration and load driving. The load drive has two parts: DC motor (rear wheel drive) and servo motor (front wheel steering). Then the final data of NodeMCU is wirelessly transmitted to the router, and then transmitted to the terminal smart device through the 4G network, thus completing the task of monitoring automatic driving. At present, artificial intelligence, the Internet of Things, and autonomous driving technologies are developing rapidly. The highlights of this research are three parts. The first part is the use of two image analysis methods to complete the automatic driving work, and the second part is the hardware architecture and circuit combination. The third part is that the system of this research is designed based on the Internet of Things architecture. The proposed system is to realize the data integration of various analysis platforms such as Raspberry Pi 4B, Jetson Nano, Motoduino U1, and mobile phone APP. The proposed method can not only realize the driving of an autonomous driving model car through image analysis but also monitor the autonomous driving process.

---

Keywords

---