Location Inference Pipeline for Noisy Visit Recognition

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing location tracking systems suffer from noisy user and place data due to poor reception, sensor limitations, and incomplete databases, leading to inaccurate location identification and visit recognition.

Innovation Solution

An inference pipeline system that utilizes validated location data to recognize user visits, estimate visit probabilities, and generate user location profiles by combining sensor data, place attributes, and external factors, using machine learning to refine predictions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If GPS, Wi-Fi, and cellular signals are used to track location, then location tracking coverage is improved, but data accuracy deteriorates due to poor reception and sensor limitations

Engineering Contradiction:

Improvelocation tracking coverageVSAvoidlocation data accuracy

Core Design Contradiction:

Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The patent combines multiple location data sources (GPS, Wi-Fi, cellular signals, sensor data) into a unified location estimate. By merging data from these diverse sources, the system achieves both broad coverage and improved accuracy, as each source compensates for the weaknesses of others.

Inventive Principle:

Principle #5Merging (Combining)

Solution Approach 2:

The system uses feedback mechanisms where location estimates are continuously refined based on new sensor data and comparisons with expected location patterns. This iterative refinement process improves accuracy over time by correcting errors in individual measurements.

Inventive Principle:

Principle #23Feedback

2Loss of information

If comprehensive place databases are created to improve place data coverage, then place information availability is improved, but data quality deteriorates due to incomplete and inaccurate location information

Engineering Contradiction:

Improveplace data coverageVSAvoidplace data accuracy

Core Design Contradiction:

Loss of informationVSReliability

Solution Approach 1:

The system performs preliminary validation and filtering of place data during the data collection and ingestion phase. By pre-processing place information to verify accuracy and completeness before full integration, the system maintains both comprehensive coverage and high data quality.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent replaces traditional mechanical database entry methods with automated inference algorithms that use sensor data and pattern recognition to validate and enrich place information. This substitution enables automated quality assurance at scale.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Productivity

If raw location data is used directly for visit recognition, then processing speed is improved, but identification accuracy deteriorates due to noisy data

Engineering Contradiction:

Improvevisit recognition speedVSAvoidlocation identification accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The patent segments the location recognition process into distinct stages: data collection, noise filtering, pattern matching, and visit confirmation. This segmentation allows efficient parallel processing while ensuring accuracy checks are performed at each stage, maintaining both speed and precision.

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS20250310718A1Inference pipeline system and method

Publication Date: 2025.10.02 FOURSQUARE LABS INC

AI summary

A system to infer place data is disclosed that receives location data collected on a user's mobile electronic device, recognizes when, where and for how long the user makes stops, generates possible places visited, and predicts the likelihood of a user to visit those places.