Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Cross-city traffic flow joint prediction method based on deep migration learning

A technology of transfer learning and urban traffic, which is applied in the field of cross-city traffic flow joint forecasting based on deep transfer learning, can solve problems such as insufficient data in target cities, and achieve the effect of improving traffic forecasting

Inactive Publication Date: 2019-08-20
NANJING UNIV OF AERONAUTICS & ASTRONAUTICS
View PDF9 Cites 18 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

By adopting the method disclosed in the present invention, the spatiotemporal correlation of traffic data flow and the knowledge of transfer learning can be effectively used to realize traffic flow prediction in data-sparse cities, solve the problem of insufficient data in target cities, and obtain higher prediction accuracy

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Cross-city traffic flow joint prediction method based on deep migration learning
  • Cross-city traffic flow joint prediction method based on deep migration learning
  • Cross-city traffic flow joint prediction method based on deep migration learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0039] The technical solution of the present invention will be further described in detail below in conjunction with the accompanying drawings.

[0040] A data-sparse urban traffic flow prediction method based on deep transfer learning such as figure 1 shown. The modeled data is input into an entire network model to generate a forecast of urban traffic flow for a period of time in the future, and the input data is historical traffic flow data for training. Specifically, the present invention constructs the following data as input:

[0041] x t : The traffic flow data at n moments before the forecast time point. x t ={x t |t=1,...n}

[0042] The data-sparse city traffic flow prediction method based on deep transfer learning disclosed by the present invention has the following specific process:

[0043] Step 1: Data Preprocessing

[0044] 1) Regional division: divide the city into an m*n grid map according to latitude and longitude, and call each grid a sub-region, and a...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The present invention provides a cross-city traffic flow joint prediction method based on deep migration learning, the method comprises three parts: firstly, dividing a city into regions, using a heatmap to represent city traffic, and abstracting traffic data at different time points into "image frames"; then using a ConvLSTM method to learn spatial dependence and time dependence of the data; andfinally, using condition maximum average difference to measure distribution distance of different spatio-temporal data. The method achieves the purpose of knowledge migration by introducing the condition maximum average difference to reduce difference between a source domain and a target domain. The method proposes to use the ConvLSTM and the conditional maximum average difference for the first time, uses an idea of migration learning to apply to urban traffic flow prediction, solves the problem of prediction difficulty caused by insufficient data in similar or related cities, improves accuracy of prediction and is of great significance to construction of smart transportation system, transportation planning and risk prediction for future cities.

Description

technical field [0001] The present invention proposes a cross-city traffic flow joint prediction method based on deep transfer learning, which relates to the field of intelligent transportation, and is mainly used for traffic flow prediction in cities with similar or related distributions but insufficient research data, and in the construction of intelligent transportation systems , urban traffic planning, traffic risk prediction, and command traffic dispatching are of great significance. Background technique [0002] Mobility is the pulse of cities, affecting the daily lives of millions of people. The application of a new generation of information technology enables humans to manage production and living conditions in a more refined and dynamic manner, by embedding and equipping sensors into power supply systems, water supply systems, transportation systems, buildings and oil and gas pipelines in every corner of the world Among the various objects in the production and liv...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G08G1/01G06N3/04G06N3/08
CPCG08G1/0129G08G1/0133G06N3/08G06N3/044G06N3/045
Inventor 王森章尹成语缪浩
Owner NANJING UNIV OF AERONAUTICS & ASTRONAUTICS
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products