Adaptive posture correction path planning method and system adaptive to continuous excavation TBM

By integrating deep reinforcement learning with multi-objective optimization algorithms, the problem of insufficient accuracy in tunnel boring machine trajectory correction was solved, achieving efficient and accurate correction path planning, adapting to complex dynamic construction environments, and improving the safety and efficiency of tunnel boring machine construction.

CN122431111APending Publication Date: 2026-07-21HUAZHONG UNIV OF SCI & TECH
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
HUAZHONG UNIV OF SCI & TECH
Filing Date
2026-04-13
Publication Date
2026-07-21

AI Technical Summary

Technical Problem

Existing methods for correcting the trajectory of tunnel boring machines (TBMs) are not accurate enough in complex dynamic environments, rely on manual operation, and consume a lot of computational resources. They are difficult to handle complex dynamic construction scenarios, and there is an urgent need for more efficient and accurate path planning methods for correction.

Method used

By employing a fusion approach of deep reinforcement learning and multi-objective optimization algorithms, and analyzing the multiple constraints of the tunnel boring machine, a smooth curvature correction path is established. Combining the reinforcement learning algorithm environment and the optimization objective to design a reward function, the TD3-NSGA-II algorithm is developed for global optimization to optimize the correction path.

Benefits of technology

It achieves high-precision and high-speed path planning for correction, significantly reducing position and angle errors under four correction scenarios. It has good generalization and adaptability, overcoming the limitations of traditional methods.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122431111A_ABST
    Figure CN122431111A_ABST
Patent Text Reader

Abstract

A self-adaptive posture correction planning method and system suitable for a continuous tunneling shield machine, the method comprising: calculating multiple constraint limits of a shield tail gap, a hinged angle, an overbreak volume and an inside-outside side thrust cylinder stroke difference of the continuous tunneling shield machine under a composite stratum; reasonably designing a unit stroke of curvature smoothing by using a relaxation curve theory; and deducing a recursive relationship of the designed correction path unit stroke in combination with geometric analysis, so as to obtain an overall correction path. A fusion algorithm TD3-NSGA-II is designed, a multi-objective optimization algorithm is driven by deep reinforcement learning to seek optimization, and the global search capability and dynamic adaptability of the correction path optimization are enhanced. In combination with a variety of typical trajectory deviation scenes, experimental research is carried out to verify the feasibility and effectiveness of the proposed method. The fusion method based on deep reinforcement learning and multi-objective optimization proposed in the application realizes self-adaptive planning of the trajectory correction path of the continuous tunneling shield machine, and provides protection for shield tunnel construction.
Need to check novelty before this filing date? Find Prior Art