Predictive Trajectory Selection Using Uncertainty for Safe Measurements
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining trajectories for measurements on technical systems do not effectively prioritize safety and information gain, leading to inefficient test planning and potential damage to the systems.
Innovation Solution
A computer-implemented method that selects the most informative trajectories based on predictive models with uncertainty measures, such as non-linear networks or Gaussian processes, to identify trajectories with greater uncertainty, thereby prioritizing safe and informative measurements, and iteratively trains the predictive model using these trajectories to avoid system damage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If traditional measurement trajectories are selected without uncertainty consideration, then the test planning is simpler, but the information gain is reduced and system damage risk increases
Solution Approach 1:
The patent replaces traditional mechanical trial-and-error test planning with a computational system using predictive models (Gaussian processes, neural networks) to calculate uncertainty measures and automatically select optimal trajectories, substituting computational intelligence for manual test design
Solution Approach 2:
The system implements feedback by continuously updating the predictive model with new measurement data from executed trajectories, recalculating uncertainty measures, and using this feedback to iteratively improve trajectory selection and reduce information loss
2Loss of information
If more trajectories are tested to increase information gain, then the information content increases, but the risk of system damage increases
Solution Approach 1:
The system performs preliminary action by using the predictive model to evaluate and rank trajectories based on uncertainty measures before actual measurements are taken, selecting only those trajectories that maximize information gain while staying within safety boundaries
Solution Approach 2:
The predictive model acts as an intermediary between the measurement system and the technical system, mediating trajectory selection by filtering out dangerous trajectories and selecting only safe, informative ones based on calculated uncertainty measures
3Measurement precision
If complex uncertainty measures are used to improve trajectory selection accuracy, then the measurement precision increases, but the computing resources required increase
Solution Approach 1:
The system employs multiple uncertainty measure parameters (predictive variance, expected improvement, upper confidence bound) and can switch between them or combine them based on computational constraints, allowing flexible adjustment of measurement precision versus computing resource usage
Data Source
AI summary
A computer-implemented method and device for determining trajectories from a set of trajectories for measurements on a technical system. A predictive model of the technical system includes a measure of uncertainty of the prediction of the predictive model, wherein the measure depends on trajectories from the set, wherein the trajectories from the set for which the measure indicates a greater or equal uncertainty than the measure for others of the trajectories from the set are determined.