Sensor Data Weighting for Faster Classification With Less Processing

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for processing sensor data are inefficient and require processing large amounts of data, which can be time-consuming and resource-intensive.

Innovation Solution

The method involves dividing sensor data into parts, mapping each part onto a representation (such as a tensor), assigning weights based on heuristic knowledge, and using these weights to reduce the data processed, incorporating Fourier coefficients and principal component analysis to characterize information content, and employing an artificial neural network for classification or regression.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If all parts of sensor data are processed, then classification accuracy is improved, but processing time increases

Engineering Contradiction:
Improveclassification accuracyVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent divides sensor data into multiple parts or channels, allowing selective processing of only those parts that contain relevant information. This segmentation enables the system to process a subset of data rather than all data, reducing processing time while maintaining classification accuracy through targeted analysis of informative segments.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent assigns different weights to different parts of sensor data based on their information content, treating each part with differentiated quality assessment. By identifying and prioritizing locally informative regions or channels while downweighting or excluding redundant parts, the system achieves accurate classification with reduced processing time.

Inventive Principle:
Principle #3Local quality

2Loss of information

If all parts of sensor data are processed, then information completeness is improved, but resource consumption increases

Engineering Contradiction:
Improveinformation completenessVSAvoidresource consumption
Core Design Contradiction:
Loss of informationVSUse of energy by moving object

Solution Approach 1:

The patent processes only a partial subset of sensor data parts that are deemed most informative, rather than exhaustively processing all available data. By applying partial action selectively to high-value data segments, the system maintains sufficient information completeness for accurate classification while significantly reducing computational resource consumption.

Inventive Principle:
Principle #16Partial or excessive action

Solution Approach 2:

The patent transforms sensor data into different representations (such as Fourier coefficients) and uses these transformed parameters to assess information content. By changing the representation parameters and selecting parts with highest information density, the system achieves efficient processing with reduced resource consumption while preserving essential information.

Inventive Principle:
Principle #35Parameter changes

3Productivity

If heuristic knowledge is incorporated into weight assignment, then processing efficiency is improved, but system complexity increases

Engineering Contradiction:
Improveprocessing efficiencyVSAvoidsystem complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent enables the system to automatically assess information content of different data parts and assign weights autonomously without requiring complex external control mechanisms. The self-service approach uses intrinsic properties of the data (such as entropy calculations from Fourier coefficients) to drive the weighting process, improving processing efficiency while keeping the added complexity manageable and self-contained.

Inventive Principle:
Principle #25Self-service

Data Source

PatentEP4685672A1Device and computer-implemented method for processing sensor or data
Publication Date: 2026.01.28 ROBERT BOSCH GMBH
  • EP4685672A1 patent drawingFigure 1~2
  • EP4685672A1 patent drawingFigure 3
  • EP4685672A1 patent drawingFigure 4

AI summary

A device and a computer-implemented method for processing sensor data, wherein the sensor data is divided into parts (202) and the parts of the sensor data are each mapped onto a representation, in particular a tensor (204), wherein for each representation a weight assigned to the representation is determined depending on the representation (208), which characterizes an information content of the part of the sensor data represented by the representation, wherein weights are drawn from a distribution of the weights determined for the representations (210), wherein a classification and/or regression of the sensor data is determined depending on the representations assigned to the weights drawn (212).