Decoupling Accelerometer Signals for Anomaly Detection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing vehicle event recorders face challenges in accurately recognizing acceleration signals indicating anomalous events due to the complexity of acceleration data from multiple sources, which often results in misclassification of events like rough roads as collisions or braking.

Innovation Solution

A system that decouples accelerometer signals into coarse acceleration and perturbation acceleration using coarse behavior data, allowing for the separation of normal vehicle operation from external or anomalous events, enabling more accurate detection of anomalies.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If accelerometer data from multiple acceleration sources is used to detect anomalous events, then the detection sensitivity is improved, but the measurement precision deteriorates due to signal mixing

Engineering Contradiction:
Improveacceleration signal recognition accuracyVSAvoidacceleration data complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the mixed acceleration signal into distinct components: coarse acceleration (from vehicle maneuvers) and perturbation acceleration (from external events). This is achieved through signal processing that separates low-frequency coarse movements from high-frequency perturbations, allowing independent analysis of each component to improve detection accuracy while managing data complexity

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent extracts the perturbation acceleration component from the total acceleration signal by removing the coarse acceleration component. This extraction process isolates the anomalous event signals from normal vehicle operation, enabling focused analysis on potentially dangerous events without interference from routine maneuvers

Inventive Principle:
Principle #2Taking out (Extraction)

2Reliability

If lower G-force thresholds are used to detect subtle anomalies, then the detection sensitivity is improved, but false positives increase due to rough road conditions

Engineering Contradiction:
Improveanomaly detection accuracyVSAvoidfalse positives from rough roads
Core Design Contradiction:
ReliabilityVSObject-affected harmful factors

Solution Approach 1:

The patent applies different analysis methods to different components of the acceleration signal. The coarse acceleration component is analyzed for vehicle maneuver context, while the perturbation acceleration component is analyzed for anomalous events. This localized quality approach allows lower thresholds for perturbation detection without triggering false positives from coarse rough road accelerations

Inventive Principle:
Principle #3Local quality

3Measurement precision

If coarse acceleration signals are included in anomaly detection, then the overall acceleration coverage is improved, but the detection precision deteriorates due to signal masking

Engineering Contradiction:
Improveanomalous event detection accuracyVSAvoidperturbation signal masking
Core Design Contradiction:
Measurement precisionVSLoss of information

Solution Approach 1:

The patent extracts the perturbation acceleration signal by removing the coarse acceleration component from the total signal. This extraction prevents the coarse component from masking subtle perturbations, as the perturbation analysis is performed on the residual signal after coarse removal rather than on the mixed signal

Inventive Principle:
Principle #2Taking out (Extraction)

Data Source

PatentUS10072933B1Decoupling of accelerometer signals
Publication Date: 2018.09.11 LYTX INC
  • US10072933B1 patent drawing
  • US10072933B1 patent drawing
  • US10072933B1 patent drawing

AI summary

A system for decoupling accelerometer signals includes an interface and a processor. The interface is configured to receive accelerometer data and receive coarse behavior data. The processor is configured to determine a coarse acceleration based at least in part on the coarse behavior data, determine a perturbation acceleration using the coarse acceleration and the accelerometer data, and determine an anomalous event based at least in part on the perturbation acceleration.