Deep Spatiotemporal  Networks for Shared Parking Supply and its Corresponding Shareable Duration Prediction

Shared parking supply and its corresponding shareable parking duration prediction is one of the fundamental and practical issues in shared parking systems, which plays a vital role in various tasks such as shared parking spaces allocation and shared parking pricing. Such predictions are very challenging, as the shared parking supply usually shows nonlinearities and complex spatiotemporal dependencies. In this paper, the authors propose a deep-learning-based network, which consists of two components for modeling Graph Convolutional Network (GCN) module and Long Short-Term Memory (LSTM) module to predict the shared parking supply and its corresponding shareable parking duration in the shared parking system. The GCN is used to capture the spatial dependencies, then the LSTM is leveraged to extract the temporal features. Moreover, in order to capture the periodically shifted correlations, the authors divided the input into three portions-recent, daily, and weekly. The proposed model is evaluated by real-word shared parking dataset in Chengdu, China. Experimentation shows that this method outperforms other five well-known baseline methods within an acceptable time frame. Extensive additional experiments and evaluations conducted to investigate the sensitivity of the model. All the results demonstrate the effectiveness of the proposed method.


  • English

Media Info

  • Media Type: Digital/other
  • Features: Figures; References; Tables;
  • Pagination: 16p

Subject/Index Terms

Filing Info

  • Accession Number: 01764344
  • Record Type: Publication
  • Report/Paper Numbers: TRBAM-21-01711
  • Files: TRIS, TRB, ATRI
  • Created Date: Dec 23 2020 11:27AM