Deep Learning-Based Mobile Application Design for Smart Parking

doi:10.1109/ACCESS.2021.3074887

Return

Deep Learning-Based Mobile Application Design for Smart Parking

2021-01-01

Hikmet Canlı *

Sinan Toklu

DOI：10.1109/ACCESS.2021.3074887

Original

Save

View PDF

Abstract

En 中文

In the era of Internet of Things (IoT) and smart city ecosystems, there is a need for innovative smart parking systems for more sustainable cities. With the increasing number of vehicles in the cities every year, it takes more time to find parking spaces. The solution methods developed are no longer sufficient. The time that passes while waiting for a parking space in traffic carries with it problems such as energy, environmental pollution and stress. In this study, a deep learning and cloud-based new mobile smart parking application was developed to minimize the problem of searching for parking spaces. Within the application, a service has been developed based on deep learning with Long short-term memory (LSTM) to predict the parking space. Here, dynamic access is provided to the LSTM-based model previously created through the mobile device of the user, and the process of displaying the occupancy rates of the parks at the desired place is accomplished on the mobile device by entering the relevant parameters. By this means, both energy and time savings have been achieved. With the real-time car parking data collected in the city of Istanbul in Turkey, high accuracy results were obtained. In order to demonstrate the effectiveness of the model proposed, it was compared with the Support Vector Machine, Random Forest and ARIMA methods. The results have confirmed the high accuracy and reliability that was promised.

Keywords:

Deep learning

Automobiles

Support vector machines

Vehicles

Smart cities

Space vehicles

Internet of Things

Smart city

deep learning

LSTM

support vector machine

random forest

ARIMA

AI Summary

Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.