Machine learning based self-emitting line marking construction parameter optimization system and method

By integrating multi-source heterogeneous data and physical mechanisms using machine learning-based methods, a self-illuminating road marking construction parameter optimization system was constructed. This system solves the problems of reliance on human experience and data scarcity in existing technologies, enabling the generation and optimization of high-precision construction parameters and improving construction quality and efficiency.

CN121859761BActive Publication Date: 2026-07-24INST OF COMM SCI YUNNAN PROV +1
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
INST OF COMM SCI YUNNAN PROV
Filing Date
2026-03-18
Publication Date
2026-07-24

AI Technical Summary

Technical Problem

Existing methods for optimizing construction parameters of self-illuminating road markings rely on manual experience, which is inefficient and highly susceptible to subjective factors. They lack the integration of physical mechanisms such as materials science and fluid mechanics, making it difficult to achieve high-precision predictions in data-scarce or extreme environments, and thus failing to meet complex construction needs.

Method used

A machine learning-based approach was adopted to construct a multimodal construction digital twin gene sequence, integrate graph neural networks and time encoders to process multi-source heterogeneous data, embed fluid mechanics, solidification kinetics and light scattering and transport equations, train a physical information neural network construction simulator, generate candidate construction parameter combinations, screen the optimal solution through a diffusion model, and optimize the model by combining expert feedback.

Benefits of technology

It significantly improves the accuracy of construction effect prediction, maintains stable and reliable prediction performance in data-scarce or complex environments, creates new high-performance parameter combinations, enables scientific and efficient screening of multi-objective construction needs, and improves construction quality and operability.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121859761B_ABST
    Figure CN121859761B_ABST
Patent Text Reader

Abstract

The application discloses a machine learning-based self-emitting marking line construction parameter optimization system and method, relates to the technical field of road engineering, and aims to train a diffusion model to reconstruct a high-quality digital gene sequence from random noise, generate a candidate construction parameter combination, and screen optimal construction scheme parameters through a physical information neural network construction simulator under the condition of a construction target; the application embeds a physical mechanism equation into a neural network, combines a multi-source heterogeneous data fusion technology, and significantly improves construction effect prediction accuracy, so that stable and reliable prediction performance can be maintained even in a data-scarce or complex environment scene. The diffusion model is used to generate a candidate construction parameter combination, which provides an innovative path for improving self-emitting marking line construction quality. Optimal parameter screening is realized based on a comprehensive performance evaluation formula, multiple project target demands and compliance constraints are considered, the screening process is scientific and efficient, and it is ensured that the selected scheme reaches a global optimum among performance, cost and efficiency.
Need to check novelty before this filing date? Find Prior Art