AI-driven local parking availability prediction

This report addresses the challenge of finding available parking lots in Singapore. The data from the Urban Redevelopment Authority (URA) indicates that car park ratios (parking spaces per 1,000 residents) in the CBD are significantly lower compared to suburban areas. This data underscores the inher...

Full description

Saved in:
Bibliographic Details
Main Author: Ma Yiheng
Other Authors: Kong Wai-Kin, Adams
Format: Final Year Project
Language:English
Published: Nanyang Technological University 2024
Subjects:
Online Access:https://hdl.handle.net/10356/175519
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
Description
Summary:This report addresses the challenge of finding available parking lots in Singapore. The data from the Urban Redevelopment Authority (URA) indicates that car park ratios (parking spaces per 1,000 residents) in the CBD are significantly lower compared to suburban areas. This data underscores the inherent limitations of parking infrastructure in the CBD and its contribution to lower availability rates. [10] The parking lot occupancy prediction offers a solution by leveraging existing online parking information. These services can empower drivers to make informed decisions by providing them with key details during their search for the optimal parking spot. [11] This project aims to develop service that forecasting parking availability of car parks in Singapore. This service can potentially alleviate the struggles associated with finding a suitable parking space. A Singapore map with parking spots is shown to assist drivers figure out which parking place is still with available lots currently and in 30 minutes near future. As a system based on the machine learning model, comparing with one based on live real-time data, it allows users to know in advance the likelihood of finding a parking space, thus better planning their trips and reducing anxiety associated with uncertainty. In Singapore, where parking occupancy rates can vary significantly between many factors, such as the day and night, peak hours and off-peak hours. [12] The necessity for such predictions in Singapore arises from its high vehicle density and limited parking infrastructure, making real-time and predictive information crucial for efficient urban mobility. A trained model can function independently of live updates, which can be particularly advantageous in areas with poor connectivity or limited sensor coverage. [13] Moreover, the improved prediction accuracy can further benefit drivers by reducing fuel consumption and emissions as they spend less time searching for parking, contributing to a more sustainable urban environment.