Estimator Update Simulation System for Accuracy Cost Trade-off

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Manually monitoring the accuracy of estimators and determining whether to update them is complex, as it involves balancing estimation accuracy with costs, and existing methods burden operators with significant resource and cost requirements without ensuring improved accuracy.

Innovation Solution

A simulation system that receives past learned estimators and verification data, applying an update rule based on evaluation indices to simulate the suitability of updating the estimator, including relearning and updating processes, to determine optimal update frequencies and costs.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If the estimator is updated frequently to maintain accuracy, then estimation accuracy is improved, but costs (machine resources, financial costs, personnel costs) increase significantly

Engineering Contradiction:
Improveestimation accuracyVSAvoidupdate costs
Core Design Contradiction:
Measurement precisionVSLoss of energy

Solution Approach 1:

The system automatically monitors estimation accuracy and triggers updates only when accuracy degradation exceeds a threshold, creating a feedback loop that balances accuracy maintenance with cost reduction by avoiding unnecessary updates

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The system changes the update frequency parameter dynamically based on actual accuracy degradation patterns rather than using fixed frequent updates, allowing optimization of the balance between accuracy and update costs

Inventive Principle:
Principle #35Parameter changes

2Loss of energy

If the estimator update frequency is reduced to minimize costs, then update costs decrease, but estimation accuracy may deteriorate

Engineering Contradiction:
Improveupdate costsVSAvoidestimation accuracy
Core Design Contradiction:
Loss of energyVSMeasurement precision

Solution Approach 1:

The system continuously monitors accuracy metrics and automatically triggers updates when degradation is detected, ensuring accuracy is maintained without requiring predetermined frequent update schedules

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The system autonomously determines when updates are necessary based on monitored accuracy levels, eliminating the need for manual monitoring and decision-making while maintaining optimal accuracy

Inventive Principle:
Principle #25Self-service

3Adaptability or versatility

If manual monitoring and determination of estimator updates is performed, then update decisions can be customized, but operator burden and complexity increase

Engineering Contradiction:
Improveupdate decision flexibilityVSAvoidoperator burden
Core Design Contradiction:
Adaptability or versatilityVSEase of operation

Solution Approach 1:

The system automatically monitors accuracy metrics and triggers updates based on predefined criteria, eliminating the need for manual monitoring and decision-making while maintaining flexible update strategies through configurable parameters

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The system provides automated feedback on accuracy degradation and update effectiveness, enabling flexible update rule configuration without requiring continuous manual intervention

Inventive Principle:
Principle #23Feedback

4Device complexity

If updating the estimator is performed regardless of accuracy improvement, then update process is simplified, but resources and costs are wasted without ensuring improved accuracy

Engineering Contradiction:
Improveupdate process complexityVSAvoidupdate costs
Core Design Contradiction:
Device complexityVSLoss of energy

Solution Approach 1:

The system evaluates accuracy improvement before executing updates and only performs updates when beneficial, avoiding wasted resources on unnecessary updates while maintaining a relatively simple update process

Inventive Principle:
Principle #23Feedback

Data Source

PatentUS10635078B2Simulation system, simulation method, and simulation program
Publication Date: 2020.04.28 NEC CORP
  • US10635078B2 patent drawing
  • US10635078B2 patent drawing
  • US10635078B2 patent drawing

AI summary

Reception means 81 receives an estimator learned using measured data up to a point of time in the past, verification data that is measured data from the point of time onward, and an update rule prescribing whether or not the estimator needs to be updated based on an evaluation index. Simulation means 82 simulates at least one of the evaluation index of the estimator and an update result of the estimator in a predetermined period, based on the update rule and an estimation result calculated by applying the verification data of the predetermined period to the estimator in chronological order.