Open Access

ARTICLE

IOT Based Smart Parking System Using Ensemble Learning

Walaa H. Elashmawi^1,3, Ahmad Akram², Mohammed Yasser², Menna Hisham², Manar Mohammed², Noha Ihab², Ahmed Ali^4,5,*

1 Department of Computer Science, Faculty of Computers and Informatics, Suez Canal University, 4.5 Km the Ring Road, Ismailia, Ismailia, 41522, Egypt
2 Department of Computer Engineering and Software Systems, Faculty of Engineering, Ain Shams University, Al-Abbaseya, 4543070, Egypt
3 Department of Computer Science, Faculty of Computer Science, Misr International University, 28 KM Cairo–Ismailia Road, Cairo, 44971, Egypt
4 College of Computer Engineering and Sciences, Prince Sattam Bin Abdulaziz University, Al-Kharj, 11942, Saudi Arabia
5 Higher Future Institute for Specialized Technological Studies, Cairo, 3044, Egypt

* Corresponding Author: Ahmed Ali. Email:

Intelligent Automation & Soft Computing 2023, 36(3), 3637-3656. https://doi.org/10.32604/iasc.2023.035605

Received 27 August 2022; Accepted 13 December 2022; Issue published 15 March 2023

Abstract

Parking space is usually very limited in major cities, especially Cairo, leading to traffic congestion, air pollution, and driver frustration. Existing car parking systems tend to tackle parking issues in a non-digitized manner. These systems require the drivers to search for an empty parking space with no guarantee of finding any wasting time, resources, and causing unnecessary congestion. To address these issues, this paper proposes a digitized parking system with a proof-of-concept implementation that combines multiple technological concepts into one solution with the advantages of using IoT for real-time tracking of parking availability. User authentication and automated payments are handled using a quick response (QR) code on entry and exit. Some experiments were done on real data collected for six different locations in Cairo via a live popular times library. Several machine learning models were investigated in order to estimate the occupancy rate of certain places. Moreover, a clear analysis of the differences in performance is illustrated with the final model deployed being XGboost. It has achieved the most efficient results with a score of 85.7%.

---

Keywords

---