Mobile Sensor Signal Evaluation Using Hierarchical Histograms

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for processing sensor signals on mobile devices face challenges in identifying characteristic data patterns due to limited storage capacity and computing power, particularly in real-time data analysis and outlier detection, where kernel density estimation requires high computational resources and storage.

Innovation Solution

A method using a hierarchical distribution of reference data values with varying resolutions, allowing for efficient data evaluation by assigning data values to sections with relative frequencies, enabling outlier detection and classification without external resources, and updating the reference distribution with new data.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If kernel density estimation is used for sensor signal processing, then measurement precision and classification accuracy are improved, but device complexity and computational resource requirements increase significantly

Engineering Contradiction:
Improveclassification accuracyVSAvoidcomputational resource requirements
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the continuous data range into discrete bins or intervals, transforming the continuous kernel density estimation problem into a discrete histogram-based approach. This segmentation reduces computational complexity by replacing complex integration operations with simple frequency counting and arithmetic operations, making the solution feasible for mobile devices with limited resources while preserving classification accuracy

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent uses simplified histogram data structures instead of complex kernel density models, employing lightweight computational approaches that consume fewer resources. The histogram-based method uses simple arrays to store frequency counts, replacing memory-intensive kernel functions with efficient lookup tables, thereby reducing both computational and memory requirements

Inventive Principle:
Principle #27Cheap short-living objects (Disposable)

2Measurement precision

If high-resolution reference distribution is used, then measurement precision is improved, but storage capacity requirements increase

Engineering Contradiction:
Improvereference distribution resolutionVSAvoidstorage capacity
Core Design Contradiction:
Measurement precisionVSQuantity of substance

Solution Approach 1:

The patent divides the reference distribution into discrete bins with configurable resolution, allowing the system to balance between precision and storage. By segmenting the continuous distribution into manageable intervals, the patent enables efficient storage using compact histogram structures while maintaining sufficient resolution for accurate outlier detection

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent dynamically adjusts the bin resolution parameter based on available storage capacity and performance requirements. By changing the resolution parameter, the system can adapt the reference distribution to fit within memory constraints while preserving the ability to detect meaningful outliers, effectively trading off precision for storage efficiency when necessary

Inventive Principle:
Principle #35Parameter changes

3Productivity

If real-time data analysis is performed, then productivity is improved, but use of energy and computational power increase

Engineering Contradiction:
Improvereal-time processing speedVSAvoidenergy consumption
Core Design Contradiction:
ProductivityVSUse of energy by moving object

Solution Approach 1:

The patent replaces computationally intensive kernel density estimation operations with simpler arithmetic operations on histogram data. By substituting complex mathematical computations with efficient array lookups and basic arithmetic, the patent enables real-time processing on mobile devices with limited processing power and energy resources, maintaining productivity while reducing energy consumption

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentEP3598077B1Multilayer signal evaluation
Publication Date: 2020.09.02 NEXENIO GMBH
  • EP3598077B1 patent drawingFigure 1A
  • EP3598077B1 patent drawingFigure 1B
  • EP3598077B1 patent drawingFigure 2

AI summary

The invention relates to a method for processing sensor signals from a sensor (312) using a mobile device (300). The method comprises providing a reference distribution (220, 308) of reference data values ​​(202) characteristic of a physical object's status, acquiring a plurality of data values ​​(310) of the physical object from sensor signals of the sensor (312) using the mobile device (300), and evaluating the acquired data values ​​(310) using the mobile device (300). The evaluation further comprises determining an overall average probability using data-value-specific average probabilities of the acquired data values ​​(310) and checking whether the acquired data values ​​(310) are characteristic of the physical object's status.If the data values ​​(310) are characteristic of the status of the physical object, the procedure further includes generating a confirmation signal that the acquired data values ​​(310) are characteristic of the status of the physical object, and updating the reference distribution (220, 308) stored in the memory (306) of the mobile device (300) with the acquired data values ​​(310).