Parking Occupancy Research Articles

Transferrable contextual feature clusters for parking occupancy prediction

Recently, real-time parking availability prediction has attracted much attention since the rapid development of sensor technologies and urbanisation. Most existing works have applied various models to predict long and short-term parking occupancy using historical records. However, historical records are not available for many real-world scenarios, such as new urban areas, where parking lots are fast adjusted and extended. In this paper, we aim to predict parking occupancy using historical data in other areas and contextual information within the targeted area that lacks historical data. We propose a two-step framework to first learn the important contextual features from areas where parking records already existed. Then we transfer these features to the other new areas without historical data records. Through conducting a real-world dataset with various clustering methods combined with different regression models, we observe that multiple contextual features are likely to influence parking availability prediction. We find the best combination (i.e., k-shape clustering algorithm and LSTM regression model) to build parking occupancy prediction model based on the subsequent quantitative correlation analysis between contextual features and parking occupancy. The experimental results show that (1) the conventional internal clustering evaluation does not work well for spatio-temporal data clustering for the prediction purpose; (2) our proposed approach achieves approximately 3% error rate in 30 minutes of prediction, which is significantly better than the estimation of the occupancy rate using the rate in the adjacent regions (13.3%).

Satellite parking: a new method for measuring parking occupancy

Abstract. Parking management plays a critical role in keeping urban spaces accessible and urban managers strive for an optimal balance between not enough and too much parking. Deciding which parking space can be liberated or needs to be extended requires detailed data on parking occupancy trends. In person inspection and in-situ sensors can provide such data but are too costly for city wide deployment. High-resolution satellite imagery is becoming more affordable, has the advantage of instantaneously collecting information from the whole city, is continuously being updated, and available for several years now to allow building a time series. Yet, identifying cars in satellite imagery is not a trivial task. We propose a method for classifying parking spot occupancy based on thresholding the reflectance range. The method requires individual parking spot data to be available and analyses each parking zone individually. We tested the method on a 0.5 metre resolution image (Pleiades satellite) that was specifically ordered for this purpose during a clear spring day in a medium-size city. The method has the advantage of not requiring extensive training data and is non-parametric. To assess accuracy, we collected ground truth data for the exact same moment as the image was ordered. The colour bands (blue, green, and red) performed equally well, while NIR seriously underperformed. We achieved a F1 score of 0.82 for all parking spots in the ground truth. The method is sensitive to tree canopy. When removing the tree obscured spots, the F1 score increased to 0.85. Tree canopy spots were automatically determined and filtered using NDVI.

A machine learning and simulation-based dynamic parking choice model for airports

Parking choice models have been developed in previous works with certain limitations, such as fixed demand, lack of real-time knowledge of parking availability at the parking facilities, and fixed parking duration, to name a few. Furthermore, such models when developed using mathematical optimization approaches, suffer from computational complexity that limits their practical applicability. This paper proposes a machine learning and simulation-based dynamic parking choice model for airport parking. It considers real-time variable parking demand, which influences access time and parking cost to various parking lots. Drivers ultimately decide to seek a parking lot depending on: (1) their value of time; (2) parking duration, and (3) availability of a parking lot. The model uses a Random Forest classifier to predict a driver's choice of a parking lot. The results, based on some empirical data, show that parking demand is highly correlated with a driver's decision to choose a parking lot, followed by parking duration and the driver's value of time. The model's accuracy in predicting a driver's parking choice is found to be 99.6%. The model provides real-time parking occupancy and can be very useful for managing airport parking. The model can be used for seeking a parking space by connected vehicles enabled with real-time information on parking availability at various parking locations (or garages). Future works may include extending the method for autonomous vehicles parking allocation and building a user-friendly dashboard with real-time traffic information for automated ground vehicle systems.

Real-Time Parking Using AI & ML

Real time parking systems have a potential to reduce the congestion in crowded areas by providing real time indications of occupancy of parking spaces. To date, search systems are mostly implemented for indoor environments using costly sensor based techniques. Consequently, with the increasing demand for parking systems in outdoor environments, inexpensive image based detection methods have become a focus of research and development recently. Motivated by the remarkable performance of Convolutional Neural Networks (CNNs) in various image category recognition tasks, this study presents a robust parking occupancy detection framework by using a deep CNN and a binary Support Vector Machine (SVM) classifier to detect the occupancy of outdoor parking spaces images. The classifier was trained and tested why the features learned by the deep CNN (convolutional neural network) from public datasets (PKlot) having different illuminance and weather conditions. The indicates the great potential of this method to provide a low- cost and reliable solution to the parking systems in outdoor environments.

Spatiotemporal Clustering of Parking Lots at the City Level for Efficiently Sharing Occupancy Forecasting Models.

This study aims to address the challenge of developing accurate and efficient parking occupancy forecasting models at the city level for autonomous vehicles. Although deep learning techniques have been successfully employed to develop such models for individual parking lots, it is a resource-intensive process that requires significant amounts of time and data for each parking lot. To overcome this challenge, we propose a novel two-step clustering technique that groups parking lots based on their spatiotemporal patterns. By identifying the relevant spatial and temporal characteristics of each parking lot (parking profile) and grouping them accordingly, our approach allows for the development of accurate occupancy forecasting models for a set of parking lots, thereby reducing computational costs and improving model transferability. Our models were built and evaluated using real-time parking data. The obtained correlation rates of 86% for the spatial dimension, 96% for the temporal one, and 92% for both demonstrate the effectiveness of the proposed strategy in reducing model deployment costs while improving model applicability and transfer learning across parking lots.

Model free Reinforcement Learning to determine pricing policy for car parking lots

Finding a parking space has not only become painful but also costs a lot in most of the metropolitan cities. With the increase in number of vehicles and limited resources such as manpower and space, the need for effective management of parking lots has increased. Improper management of parking lots can have negative consequences such as traffic congestion, wastage of time in search of parking spaces, air pollution and even loss of revenue for the parking lot managers. Dynamic pricing is a powerful tool to control the behavior of drivers by diverting them towards the unoccupied and cheaper parking lots. Though there are several existing dynamic pricing strategies, determining the right prices is quite challenging due to lack of knowledge of drivers’ behavior and several uncertainties like harsh weather and special days. In this paper Reinforcement Learning(RL) technique called Q-learning is used to calculate the dynamic prices for parking lots on hourly basis without the need of prior information about the system. Crucial factors like distance of the parking lots from the city centers, weather and holidays are considered in the proposed algorithm to achieve better accuracy. Price Elasticity of Demand (PED) is used in the proposed work to calculate the new state(occupancy) when an action(dynamic price charged by the parking lot owner) is taken place. Hourly prices are estimated using the proposed algorithm and simulation results show that the calculated prices can efficiently manage parking occupancy during peak and off peak hours. The simulation output also shows that the proposed algorithm can successfully increase the revenue of the parking lot owners.

Estimating the Measurement Uncertainty of the Number of Vehicles in a Car Park Using an Indirect Method

The main problem with counting vehicles in rest areas in Europe and America is the overflowing of parking lots in the rest areas. Thus it is impossible to count parked vehicles with traditional, direct methods, which use cameras and Lidar (light detection and ranging) to detect the presence of vehicles in individual parking spots. The solution to this problem may be an indirect method which uses cameras to count the vehicles entering and leaving the rest area and which sorts the vehicles into categories. This article introduces a method for determining car park occupancy in rest areas using indirect measurement and evaluates the uncertainty of this method for determining the occupancy. This indirect method counts the vehicles entering and leaving the car park and sorts the vehicles into categories. The difference between the number of vehicles that entered and the number of vehicles that left in a given time gives the number of vehicles occupying the car park, i.e., the parking space demand for a given time. This demand will vary over time. In order to register the vehicles entering and leaving, mains-free supply video cameras were installed next to the entrance and exit. The counting and categorizing could be conducted organoleptically, semi-automatically, or using an image computer analysis using artificial intelligence networks. Because the difference between the number of vehicles entering and leaving is calculated over a long period, a relative error (regarding car park capacity) might be grave even when the counting errors are minor. The authors will show how the certainty of this indirect counting method can be improved.

Deep learning-based parking occupancy detection framework using ResNet and VGG-16

Deep learning-based parking occupancy detection framework using ResNet and VGG-16

PREDICTION OF PARKING SPACE AVAILABILITY USING ARIMA AND NEURAL NETWORKS

It may be critical for drivers to have information about the occupancy rates of the parking spaces around their destination in order to reduce the traffic density, a non-negligible part of which caused by the trips to find an available parking space. In this study, we predict parking occupancy rates (and thus, space availability) using three different techniques: (i) auto-regressive integrated moving average model, (ii) seasonal auto-regressive integrated moving average model and (iii) neural networks. In the implementation phase, we use the data set of the on-street parking spaces of the well-known “SFpark” project carried out in San Francisco. We take into account not only the past occupancy rates of parking spaces, but also exogenous variables that affect the corresponding occupancy rates as day type and time period of the day. We make predictions with different model structures of each of the considered methods for each parking space with different parking occupancy patterns in the data set and then compare the results to find the best model design for each parking space. We also, evaluate the results in terms of the superiority of the methods over each other and note that the performance of neural networks is better than those of the other approaches in terms of the mean squared errors.

Spatio-temporal parking occupancy forecasting integrating parking sensing records and street-level images

The prediction of parking occupancy is of great importance to urban planning. As the number of cars increases and parking resources become limited, the lack of parking supply has become a challenge for urban design. Previous works ignore the correlation between car parks when predicting parking occupancy, which limits the accuracy of the prediction. To address this issue, this study proposes a Temporal-GCN-based correlated parking prediction model (CPPM) to forecast the temporal occupancy of car parks. In particular, the model utilises Convolutional Neural Networks (CNN) and Bayesian probabilities to extract street view similarities in car parks, as well as their spatial correlations, cosine similarity is used to calculate the activity type similarity, and Graph Convolutional Networks (GCN) and Gate Recurrent Units (GRU) are integrated to predict spatio-temporal car park occupancy, taking into account both temporal parking records, similarities in car parks, and their spatial correlations. We conducted two case studies in Ningbo and Beijing, China, integrating over 10 million parking sensing records and corresponding street view images of parking lots to predict parking occupancy. The results show that our model has outstanding performance over the baselines and can be extended for various types of car parks in cities of different sizes and different levels of development. The results also reveal the parking preferences of the citizens of Ningbo and Beijing, which is valuable for a quantitative understanding of commuters’ parking patterns and behaviour and can be used as a guide for urban planning and management.

An Integrated Approach for the Near Real-Time Parking Occupancy Prediction

In a city, the usage optimization of parking spaces with a near real-time response to car drivers can significantly reduce the unnecessary cruising for parking and the additional congestion of regional traffic. As the foundation to achieve such an optimization, a parking occupancy prediction method is required to address the emerging challenges of training a simple but effective model. To fill the gap, this paper proposes a novel approach that enables the integration of Time Series Decomposition (TSD), Gated Recurrent Unit (GRU), and First-order Model-agnostic Meta-learning (FOMAML) for feature engineering, model building, and model pre-training, respectively. Moreover, as shown by a detailed evaluation, such an integration strengthens the proposed approach, named Meta TSD-GRU, which outperforms other state-of-the-art methods with 1) prediction errors reduced by about 45% on average, 2) the speed of model adaptation and convergence improved about 2 and 102 times against the methods with and without pre-training, respectively, and 3) the generalizability of the model enhanced to handle various time intervals of forecasting and types of parking lots under a consistent and stable performance.

Tracking Parking Search and Occupancy in Zurich

This article describes how parking search times and occupancy in a highly attractive area in central Zurich (Switzerland) develop on a weekday morning when a farmer’s market takes place on Bürkliplatz, a central market place inside the study area. Individual vehicles were tracked by bike, drivers were interviewed about their parking behaviour and their route was tracked using GPS. Additionally, the parking occupancy was registered every fifteen minutes during a five-hour time period. A connection between market opening hours and parking dynamics within the perimeter was observed. Drivers usually overestimate their parking search duration.

Managing parking with progressive pricing

Parking supply and demand are often imbalanced in urban areas, causing adverse consequences such as excessive search times and long walking distances. Many parking authorities price parking as a demand management strategy by charging either a fixed daily fee or an hourly price for parking. An emerging alternative is progressive pricing, whereby drivers pay an hourly price that increases if their tracked parking duration is longer than a predetermined threshold. This study investigates the optimal design of progressive pricing for revenue and social welfare maximization when there are two market segments. We study equilibrium properties of progressive pricing and show that its optimal design for revenue maximization segmentizes the demand. In contrast, progressive pricing does not improve social welfare over hourly pricing, because any additional surplus accrued by drivers is offset by the revenue of the parking authority. We develop a micro-simulation model of the City of Toronto’s downtown core with parking capability, and show that progressive pricing lowers average parking occupancy and search time in high demand parking clusters by 5.6% and 12.5%, respectively, compared to hourly pricing.

Parking occupancy prediction and analysis - a comprehensive study

The seemingly interminable search for an available parking spot wastes time and fuel while polluting the environment. Yet, the inefficiency of the process persists, due to a lack of real-time and near-future information. Researchers are addressing this issue by developing systems that combine various modeling techniques with historical and real-time data to predict parking occupancy. The objective of this study was to provide a comprehensive review of the existing academic research on this topic by conducting a thorough examination of 108 published research papers, journals, and articles. The findings revealed that hybrid models that combine different approaches are most effective in making accurate predictions, as they capture a wider range of patterns and relationships in parking occupancy dynamics. Neural network models, particularly convolutional and recurrent neural networks, demonstrate exceptional capabilities in processing intricate spatial and temporal patterns. Deep learning models, particularly long short-term memory models, excel in handling large and complex datasets and show potential for predicting short-term parking availability. Time series models, such as autoregressive integrated moving average models, are deemed suitable for capturing temporal patterns. The findings of the study will serve as a valuable resource for researchers who can build upon the existing knowledge and make more informed decisions when developing parking prediction models, ultimately contributing to the advancement of the field.

Examining parking occupancy prediction models: a comparative analysis

Parking is a persistent challenge in urban areas, especially on college campuses, as limited parking spaces and increasing vehicular traffic are at the forefront of campus administrators’ concerns. Lack of parking availability leads to frustration among students, faculty, and staff and results in increased traffic congestion, environmental pollution, and reduced overall productivity. Accurate prediction of parking occupancy plays a vital role in optimizing parking resource allocation, minimizing congestion, and improving the overall campus experience. Thus, this study comprehensively analyzes parking occupancy predictions for a college campus garage, using several different models. The dataset spans from January 2022 to March 2023, providing insights into the models' performance for more than one year. The research aims to utilize random forest, decision tree, linear regression, and support vector regression (SVR) models and compare their effectiveness, using evaluation metrics such as root mean square error (RMSE), mean absolute error (MAE), and R-squared (R2). Random forest outperformed decision tree, SVR and linear regression models in terms of accuracy, as indicated by the lower MAE values. Additionally, random forest achieved a better classification performance, and a higher R2 suggests a stronger correlation between the predicted and actual occupancy values. Conversely, SVR displayed the weakest performance among the four models, with a high MAE and negative R2 value. The findings of this research will aid in improving parking management systems and contribute to the development of efficient and sustainable parking solutions on college campuses and in urban areas.

Analysing and Identifying Geospatial Key Factors in Smart Cities – Model Enhancements in the Use Case of Carpark Occupancy

Urban planning benefits significantly from improved knowledge concerning spatiotemporal relationships and patterns in cities based on geospatial factors. In this context, the potential of geodata seems inexhaustible. Applications include limited-service offers like carparks, the occupancy of which is controlled by geospatial factors characterized by their spatiotemporal patterns. This paper proposes an enhanced model for identifying geospatial key factors, tying in with an existing geo-analytics model. Our approach combines real-world empirical data for off-street parking with open-source geodata on points of interest. We formulate stabilization measures in different model-enhancement stages to optimize model reliability and fit, based on analyses of statistical characteristics. Additionally, we consider modifying the choice of geospatial factors in order to reduce multicollinearity. Our results show improved reliability of geo-analytics for the identification of urban spatiotemporal relationships.

Coins, Cards, or Apps: Impact of Payment Methods on Street Parking Occupancy and Wait Times

Coins, Cards, or Apps: Impact of Payment Methods on Street Parking Occupancy and Wait Times

Using a Fuzzy Logic Integrated Machine Learning Algorithm for Information Fusion in Smart Parking

The free flow of people and products within metropolitan areas depends on well-managed transportation systems. However, public parking places in smart cities are often limited by traffic, causing cars and residents to waste time, money, and fuel. To counteract this issue, today's automobile systems combine information fusion with intelligent parking solutions. In this research, we present a Fuzzy Logic Integrated Machine Learning Algorithm (FL-MLA) for use in smart parking and traffic management in a metropolis. The FL-MLA use fuzzy induction to distinguish between parked and moving vehicles while calculating traffic flow. The suggested technique efficiently resolves the problem of locating suitable parking places by avoiding incorrect configurations that govern traffic management difficulties. Therefore, the FL-MLA is used in traffic management systems to boost performance metrics like efficiency ratio (98.1%) and accident detection (98.1%) based on simulation results like reduced energy consumption (95.3%), more accurate traffic estimation (97.9%), higher average daily park occupancy (97.2%), and higher efficiency ratio (98.1%).

Smart parking occupancy detection method using smart app

Smart parking occupancy detection method using smart app

Coins, Cards, or Apps: Impact of Payment Methods on Street Parking Occupancy and Wait Times

Coins, Cards, or Apps: Impact of Payment Methods on Street Parking Occupancy and Wait Times