Vehicle Data Sampling Feedback for Adaptive Anomaly Detection

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing data collection methods from vehicles suffer from suboptimal sampling criteria that may include irrelevant data and miss relevant data, leading to inefficient error detection, and anomaly detection tools are not adaptable to unknown situations.

Innovation Solution

A process that optimizes data collection by using mutual reinforcement feedback between sampling and anomaly detection, adjusting sampling criteria to focus on likely anomalies and refining anomaly detection logic based on statistical models and machine learning.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of energy

If sampling criteria are used to collect data from a portion of vehicles, then the total cost of data collection is reduced, but the sampling criteria may include too much irrelevant data or omit the most relevant data

Engineering Contradiction:

Improvecost of data collectionVSAvoidaccuracy of error detection

Core Design Contradiction:

Loss of energyVSMeasurement precision

Solution Approach 1:

The system implements feedback loops where anomaly detection results are fed back to update sampling criteria. The statistical model continuously learns from detected anomalies and adjusts future sampling decisions, creating a self-improving system that reduces irrelevant data collection while maintaining high detection accuracy

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The sampling criteria dynamically change parameters such as sampling rate, selection probability, and data sources based on statistical analysis of received data and detected anomalies. This allows the system to adapt to different vehicle types, operating conditions, and error patterns, optimizing the balance between data collection cost and detection accuracy

Inventive Principle:

Principle #35Parameter changes

2Ease of operation

If anomaly detection tools use pre-set logic, then the detection process is straightforward, but the tools may miss errors since they are not adaptable to unknown situations

Engineering Contradiction:

Improvesimplicity of anomaly detectionVSAvoidadaptability to unknown situations

Core Design Contradiction:

Ease of operationVSAdaptability or versatility

Solution Approach 1:

The anomaly detection system transitions from static pre-set logic to dynamic adaptive detection. The statistical model continuously updates its parameters and detection thresholds based on incoming data and identified anomalies, allowing the system to adapt to new error patterns and unknown situations while maintaining operational simplicity through automated adjustments

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system performs self-learning and self-adjustment through the statistical model that automatically improves anomaly detection capabilities without external intervention. The model learns from detected anomalies and automatically refines detection logic, making the system both simple to operate and highly adaptable to unknown situations

Inventive Principle:

Principle #25Self-service

3Reliability

If data is collected from every single vehicle, then complete coverage is achieved, but the cost and complexity of data collection increases significantly

Engineering Contradiction:

Improvecompleteness of error detectionVSAvoidcomplexity of data collection system

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The system extracts and focuses on the most valuable data by using statistical models to identify and prioritize vehicles and data points with highest anomaly probability. Instead of collecting data from all vehicles equally, the system extracts only the critical subset that provides maximum detection value, reducing system complexity while maintaining reliability

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The system applies partial action by collecting data from a strategically selected portion of vehicles rather than all vehicles. The sampling criteria are designed to capture sufficient statistical information for reliable error detection while avoiding the excessive complexity of universal data collection

Inventive Principle:

Principle #16Partial or excessive action

4Productivity

If sampling criteria are adjusted to focus on specific vehicles, then relevant data collection efficiency improves, but the criteria may become too restrictive and miss other important anomalies

Engineering Contradiction:

Improveefficiency of data collectionVSAvoidcompleteness of anomaly detection

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The statistical model serves multiple functions simultaneously: it analyzes received data, detects anomalies, updates sampling criteria, and validates detection results. This multi-functional approach allows the system to maintain both efficient targeted sampling and comprehensive anomaly detection coverage through a single integrated system

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentUS20250371642A1Systems and methods for data collection from vehicles using sampling and anomaly detection

Publication Date: 2025.12.04 TOYOTA JIDOSHA KK

AI summary

Provided are a method, system, and device for optimizing data collection from a plurality of vehicles. The method may include receiving data collected from the plurality of vehicles selected based on one or more sampling criteria; generating a statistical model based on the received data; detecting, based on the statistical model, anomalies in the received data; and updating the one or more sampling criteria based on the anomalies in the received data.