Signal Evaluation Using Detrended Fluctuation Analysis

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional methods for air pollution forecasting in urban areas, particularly from road traffic, often rely on weather conditions and fail to accurately account for temporal fluctuations in immission measurements, leading to inadequate or premature measures.

Innovation Solution

A method utilizing Detrended Fluctuation Analysis (DFA) to analyze time series data of immissions, allowing for the identification of long-term correlations and short-term changes, enabling more precise and localized forecasting and countermeasure implementation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional forecasting methods relying on weather conditions are used, then the forecasting process is simple, but the accuracy of emission forecasts deteriorates

Engineering Contradiction:
Improveaccuracy of emission forecastsVSAvoidcomplexity of forecasting method
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent applies preliminary action by calculating the fluctuation exponent from historical immission time series data in advance. This pre-computed exponent characterizes the temporal fluctuation pattern and is stored for later use in forecasting. When forecasting is needed, this pre-calculated exponent is combined with current weather forecasts to generate emission predictions, avoiding the need to re-analyze historical data each time and improving both accuracy and efficiency.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The fluctuation exponent serves as an intermediary parameter that bridges historical immission data and future emission forecasts. Instead of directly using complex historical time series or simple weather conditions, the patent introduces this intermediate metric that captures the essential temporal fluctuation characteristics. This intermediary enables the combination of historical patterns with weather predictions to produce accurate forecasts without requiring direct analysis of the full historical dataset.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Measurement precision

If hourly or daily mean values are calculated for monitoring purposes, then the monitoring process is simple, but the temporal resolution of immission measurements deteriorates

Engineering Contradiction:
Improvetemporal resolution of immission measurementsVSAvoidcomplexity of data processing
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent extracts the fluctuation exponent as a specific characteristic from the complex immission time series data. Instead of processing the entire detailed time series or calculating simple mean values, it extracts this single parameter that captures the essential temporal fluctuation pattern. This extracted exponent can then be used for forecasting without needing to retain or process the full high-resolution historical data, thus maintaining temporal resolution benefits while simplifying storage and computation.

Inventive Principle:
Principle #2Taking out (Extraction)

3Area of stationary object

If decentralized distributed solutions are used to measure larger areas, then the spatial coverage is improved, but the complexity of evaluation increases

Engineering Contradiction:
Improvespatial coverage of measurementsVSAvoidcomplexity of evaluation
Core Design Contradiction:
Area of stationary objectVSDevice complexity

Solution Approach 1:

The patent transforms complex spatial-temporal immission data into a simplified parameter - the fluctuation exponent - that characterizes temporal patterns. By applying this transformation to data from multiple decentralized measuring points, each location generates its own exponent that captures local temporal behavior. This parameter transformation simplifies the evaluation of decentralized measurements, allowing complex spatial patterns to be represented through comparable exponent values from different locations.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP3712786A1Method for evaluating at least one signal
Publication Date: 2020.09.23 ROBERT BOSCH GMBH
  • EP3712786A1 patent drawingFigure 1
  • EP3712786A1 patent drawingFigure 2
  • EP3712786A1 patent drawingFigure 3

AI summary

Method for evaluating at least one signal representing a time series of measured values, wherein the measured values ​​relate to emissions, wherein the at least one signal is evaluated using a trend-correcting fluctuation analysis, so that a trend-corrected signal is obtained, which in turn is evaluated to detect a temporal change in the fluctuation of the measured values ​​and to evaluate this temporal change.