A checkpoint travel time estimation method based on mixed integer optimization and spatial KNN

CN121980373BActive Publication Date: 2026-06-30SHANDONG UNIV OF SCI & TECH
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202610450258.4
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2026-04-08
Publication Date
2026-06-30
Estimated Expiration
2046-04-08

AI Technical Summary

Technical Problem

Existing checkpoint travel time estimation technologies cannot effectively identify and eliminate abnormal delay noise caused by vehicle stopping behavior, and the extrapolation results in low-traffic or blind-spot road sections without detection data lack physical meaning, resulting in inaccurate speed estimation.

Method used

A collaborative estimation method based on mixed integer optimization and spatial KNN is adopted. Through data cleaning, travel time decomposition, flow determination and spatial KNN topology constraints, the method can accurately identify abnormal dwell times and extrapolate the travel time of blind spots.

Benefits of technology

Abnormal noise was eliminated in sections with traffic, and physical spatial correlation was restored in blind spot sections, improving the accuracy and precision of speed estimation and reflecting the true dynamic operating status of the road network.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121980373B_ABST
    Figure CN121980373B_ABST
Patent Text Reader

Abstract

This invention relates to the field of traffic control systems, specifically to a checkpoint travel time estimation method based on mixed-integer optimization and spatial KNN, comprising the following steps: cleaning mapped checkpoint data, calculating initial spatiotemporal parameters, and removing anomalies based on absolute median; decomposing travel time and removing anomaly dwellers; dividing the road network according to traffic flow thresholds, estimating high-traffic road segments using mixed-integer optimization verification and Bayesian dynamic fusion, and performing parameter extrapolation based on spatial KNN topological constraints and congestion characteristics inverse solution for low-traffic road segments; fusing the results from dual-source road segments to construct a global parameter vector, and iteratively optimizing through an iterative convergence mechanism to output a high-precision global travel time field. This method utilizes absolute median to construct a robust filtering mechanism to remove extreme anomalies; it performs parameter extrapolation for low-traffic road segments, achieving high-precision estimation while ensuring physical interpretability, accurately reconstructing the dynamic operating state of the road network.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Traffic flow prediction method based on genetic algorithm optimized LSTM neural network

    CN109243172A

  • Time-sharing electric quantity prediction method based on logarithmic load density growth curve

    CN121602348A