Vehicle Sensor Data Labeling via Confidence-Scored Object Matching

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current methods for obtaining and labeling vehicle sensor data are complex, expensive, and time-consuming, especially in the context of autonomous vehicle control, where high-quality labeled data is crucial for training machine learning systems.

Innovation Solution

A computer-implemented method that processes sensor data from distance-ranging sensors on a vehicle in conjunction with object data from the environment, determining a possible association, calculating a confidence value, and labeling the respective portion of the sensor data with attribute data associated with the object, and in response to the confidence threshold, thereby generating labeled vehicle sensor data.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If human annotation or expensive camera sensors are used to label sensor data, then the quality and reliability of labeled data sets improve, but the cost and time required for obtaining labeled data increase significantly

Engineering Contradiction:

Improvequality of labeled dataVSAvoidtime required for obtaining labeled data

Core Design Contradiction:

ReliabilityVSLoss of time

Solution Approach 1:

The system uses the vehicle's own distance-ranging sensor data to automatically generate labels for sensor portions, eliminating the need for external human annotation. The processor compares sensor data with object data to autonomously create labeled datasets, making the system self-sufficient in data labeling operations.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The system creates labels by copying attribute data from object data and associating it with corresponding sensor data portions. This copying mechanism allows automatic generation of accurate labels without manual intervention, significantly reducing time and cost while maintaining data quality.

Inventive Principle:

Principle #26Copying

2Reliability

Engineering Contradiction:

Improvequality of labeled dataVSAvoidcost of obtaining labeled data

Core Design Contradiction:

ReliabilityVSEase of manufacture

Solution Approach 1:

The vehicle's onboard processor performs automatic labeling using its own sensor data and received object data, eliminating the need for expensive external annotation services or specialized camera equipment. This self-service approach dramatically reduces costs while maintaining labeling quality.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The system uses inexpensive distance-ranging sensor data as the basis for generating labels, replacing expensive camera sensors and human annotators. The approach leverages readily available, low-cost data sources to achieve the same labeling function at a fraction of the cost.

Inventive Principle:

Principle #27Cheap short-living objects (Disposable)

3Measurement precision

If manual labeling processes are used, then high-quality labeled data can be obtained, but the complexity and time consumption of the process increase

Engineering Contradiction:

Improveaccuracy of labelsVSAvoidcomplexity of labeling process

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The system merges sensor data with object data in a unified processing operation. The processor simultaneously handles both data types to determine associations and generate labels, simplifying the overall process while maintaining accuracy through integrated analysis.

Inventive Principle:

Principle #5Merging (Combining)

Solution Approach 2:

The system uses confidence values as feedback to validate label associations. By calculating confidence levels for each sensor-object association, the system automatically ensures label accuracy without manual verification, reducing both complexity and time while maintaining high measurement precision.

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS20260004598A1Systems and methods for generating labelled vehicle sensor data

Publication Date: 2026.01.01 NXP BV

AI summary

A computer-implemented method comprising: receiving, at a processor, sensor data generated by at least one distance-ranging sensor on a vehicle within an environment; receiving, at the processor, object data received at the vehicle, the object data originating from a plurality of objects within the environment, wherein the object data comprises location data indicating a respective location of each object and attribute data associated with each object; processing the data to determine a possible association between a respective portion of the sensor data and at least one object of the plurality of objects; calculating a confidence value reflecting the reliability of the possible association; if the confidence value exceeds a threshold, labelling the portion of the sensor data with at least a portion of the attribute data of the associated object(s).