An industrial robot operation risk prediction and planning method based on probabilistic twin

By using Bayesian calibration and Monte Carlo simulation of the probabilistic twin model, the distortion problem of traditional digital twin models in dynamic environments is solved, achieving high-fidelity risk prediction and intelligent path planning, and improving the robot's autonomous decision-making and operational efficiency.

CN121132683BActive Publication Date: 2026-05-26SUZHOU VOCATIONAL INSTITUTE OF INDUSTRIAL TECHNOLOGY
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
SUZHOU VOCATIONAL INSTITUTE OF INDUSTRIAL TECHNOLOGY
Filing Date
2025-10-22
Publication Date
2026-05-26

AI Technical Summary

Technical Problem

When faced with dynamic and non-ideal industrial environments, traditional digital twin models suffer from physical parameter drift, leading to model distortion. Risk prediction uncertainty lacks quantitative management, and existing risk avoidance strategies are inefficient and difficult to implement intelligent decision-making.

Method used

A probabilistic twin model is established, defining the intrinsic physical parameters as random variables with prior probability distributions. The parameter distribution is updated using a Bayesian calibration method, and the operation path is optimized using Monte Carlo simulation and information gain.

Benefits of technology

It achieves high-fidelity self-calibration of the model, improves the accuracy and reliability of risk prediction, enables it to proactively adapt to environmental changes, and enhances the autonomous decision-making ability and operational efficiency of industrial robots.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121132683B_ABST
    Figure CN121132683B_ABST
Patent Text Reader

Abstract

This invention discloses a method for predicting and planning operational risks of industrial robots based on probabilistic twins. The method includes: establishing an industrial robot model, defining at least one intrinsic physical parameter affecting the dynamic behavior of the industrial robot as a random variable with a prior probability distribution, thus forming a probabilistic twin model of the industrial robot; acquiring sensor observation data from the physical world during the industrial robot's task execution, and updating the probability distribution of the intrinsic physical parameter using a Bayesian calibration method based on the difference between the observation data and the prediction results of the probabilistic twin model; probabilistically predicting the future risks of the industrial robot executing candidate task paths based on the updated probabilistic twin model, and planning or adjusting the industrial robot's task path based on the risk prediction results; this invention enables path planning to intelligently balance safety, task completion, and active learning, significantly improving the robot's autonomous decision-making ability and operational efficiency.
Need to check novelty before this filing date? Find Prior Art