The invention relates to a roadside parking availability prediction method based on a graph neural network. The method comprises the following steps: constructing parking area graph structure data and parking occupancy time sequence data; constructing a graph convolution module and a time convolution module, and combining the graph convolution module and the time convolution module into a space-time residual convolution module; the method comprises the following steps: constructing a space-time residual image convolutional neural network, capturing space correlation information and long-time dependence information of parking data, constructing an output layer neural network, and predicting a parking occupancy rate in a certain period of time in the future based on space-time features captured by the space-time residual image convolutional neural network. The beneficial effects of the invention are that the method considers the space information of the parking data, achieves the modeling of the space information into a graph, learns the space correlation between the parking regions through the powerful capability of a graph convolutional neural network, and further improves the prediction accuracy of the parking occupancy.
本发明涉及一种基于图神经网络的路边停车可用性预测方法,包括:构建停车区域图结构数据和停车占有率时序数据;构建图卷积模块和时间卷积模块,并组合为时空残差卷积模块;构建时空残差图卷积神经网络,捕获停车数据的空间关联信息和长时依赖信息,并构建输出层神经网络,基于时空残差图卷积神经网络捕获的时空特征预测未来某一段时间的停车占有率。本发明的有益效果是:本发明考虑了停车数据的空间信息,通过将空间信息建模成图,使用图卷积神经网络的强大能力学习停车区域间的空间关联,进一步提高了停车占有率预测的准确性。
Roadside parking availability prediction method based on graph neural network
一种基于图神经网络的路边停车可用性预测方法
2023-04-18
Patent
Electronic Resource
Chinese
Roadside self-service parking device and roadside parking management method
European Patent Office | 2023
|Roadside parking space management method and system and roadside parking space management pile
European Patent Office | 2021
|