Wall Diagnostic AI Training Data From Classified Sensor Signals

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for generating training data sets for artificial intelligence in wall diagnostic devices are inadequate, leading to inefficient and inaccurate object detection and classification in walls.

Innovation Solution

A method involving a classifier module that uses a trained artificial intelligence to classify unclassified sensor data, including radar data, to generate a training data set for improved object recognition and classification, utilizing additional classification information to enhance detection and classification accuracy in three spatial dimensions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If a classifier module with artificial intelligence is used to classify unclassified sensor data, then object detection precision and classification accuracy are improved, but device complexity and training data requirements increase

Engineering Contradiction:
Improveobject detection precisionVSAvoiddevice complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent applies preliminary action by pre-training the classifier module with a training data set containing classified sensor data before actual object detection. This preprocessing step enables the AI to learn patterns and improve detection precision while maintaining operational efficiency during actual use.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent introduces an intermediary training data set that bridges the gap between raw sensor data and object classification. This intermediate structured data with labels serves as a mediator to train the classifier, improving detection accuracy without requiring complex real-time processing.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Measurement precision

If additional classification information is incorporated for three-dimensional object detection, then object classification accuracy is improved, but data processing complexity and time increase

Engineering Contradiction:
Improveobject classification accuracyVSAvoiddata processing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent performs preliminary classification by generating a training data set with pre-computed classification information including object positions, types, and characteristics. This advance preparation reduces the computational burden during actual detection, improving accuracy without excessive processing time delays.

Inventive Principle:
Principle #10Preliminary action

3Reliability

If a training data set is generated and used for classifier module training, then reliability of object detection is improved, but loss of time for data generation and training increases

Engineering Contradiction:
Improvereliability of object detectionVSAvoidtraining time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent applies preliminary action by generating the training data set in advance with classified sensor data and labels. This one-time upfront investment in data preparation and model training significantly improves detection reliability during subsequent operations, reducing the time penalty to an initial setup cost rather than continuous overhead.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP4618034A1Method for generating a training dataset
Publication Date: 2025.09.17 ROBERT BOSCH GMBH
  • EP4618034A1 patent drawingFigure 1
  • EP4618034A1 patent drawingFigure 2
  • EP4618034A1 patent drawingFigure 3

AI summary

The invention relates to a computer-implemented method (300) for generating a training data set (143) for training an artificial intelligence (125) for operating a measuring device (100), in particular a wall diagnostic device, comprising: receiving (301) unclassified sensor data (172) of at least one sensor unit (101) of a measuring device (100) by a classifier module (183); classifying (303) the unclassified sensor data (172) and providing classified sensor data (174) by executing a classifier module (183) on the unclassified sensor data (172); and adding (305) the classified sensor data (174) to a training data set (143).