A multimodal traffic flow prediction method and device, a storage medium and an electronic device
By acquiring information about external conditions of the grid and using structural causal models to predict passenger flow for taxis, buses, and subways, the shortcomings of traditional peak traffic forecasting methods have been addressed. This has enabled precise scheduling of multi-modal traffic flows, improving urban traffic efficiency and residents' travel experience.
CN122245112APending Publication Date: 2026-06-19BEIHANG UNIV
View PDF 0 Cites 0 Cited by
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- BEIHANG UNIV
- Filing Date
- 2026-03-31
- Publication Date
- 2026-06-19
AI Technical Summary
Technical Problem
Traditional peak traffic forecasting methods cannot fully cover the diverse travel modes and complex traffic flow fluctuations in modern cities, resulting in inefficient traffic scheduling and serious congestion and emissions problems.
Method used
By acquiring external condition information of the target area grid, exogenous latent variables and causal endogenous latent variables are derived layer by layer using a structural causal model to predict passenger flow of taxis, buses and subways, thus achieving multimodal traffic flow prediction.
Benefits of technology
It has improved the overall efficiency of urban transportation, reduced congestion and emissions, and enhanced the travel experience for residents.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN122245112A_ABST
Abstract
This invention proposes a multi-modal traffic flow prediction method, device, storage medium, and electronic device. It acquires external condition information for each grid in a target area at the current time step. Based on the causal endogenous latent variables of the grid at the previous time step and the external condition information at the current time step, it determines the exogenous latent variables of the grid at the current time step. The exogenous latent variables of the grid at the current time step are input into a structural causal model to obtain the causal endogenous latent variables of the grid at the current time step. Based on the causal endogenous latent variables of the grid at the current time step, the predicted passenger flow value of the grid at the current time step is determined. By deriving the exogenous latent variables, causal endogenous latent variables, and passenger flow prediction value layer by layer through external condition information, and predicting traffic conditions in subsequent time steps, it can better complete traffic scheduling, improve the overall efficiency of urban traffic, reduce congestion and emissions, and improve residents' travel experience.
Need to check novelty before this filing date? Find Prior Art