TRB Highlight: “Predicting Parking Occupancy for Residential Shared Parking Lots Using Deep Learning Methods Considering COVID-19 Policies”

Title of Paper: “Predicting Parking Occupancy for Residential Shared Parking Lots Using Deep Learning Methods Considering COVID-19 Policies”

Authors: Zhipeng Niu, Xiaowei Hu, Mahmudur Fatmi, Shouming Qi, Dongling Wu, Shi An

Description: In this study, a parking occupancy prediction model based on a policy-aware temporal convolutional network (P-TCN) is proposed. Policy dependence of non-stationary data is considered in the model with LSD differentiation and analysis to accurately identify the effect of anti-pandemic policies on parking behavior. The experimental results show that the prediction results of the proposed P-TCN model are consistent with the trend of the measured values with less deviation and have better generalization abilities compared to models such as FCN, LSTM, and TCN. Under the normalization of the COVID-19 pandemic, the parking occupancy information provided by the framework will help managers better understand residents’ parking behavior and have greater flexibility in responding to emergencies. In addition, it will provide newer and more valuable insights to explore the future development of shared parking strategies.