MAC Address Estimation Using Segmentation and Learned Models

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing methods for estimating the number of users in a facility, such as restaurants, are inaccurate when terminals have a randomization function that changes their MAC addresses frequently, leading to mixed and unpredictable MAC address data.

Innovation Solution

An information processing device and method that collects MAC addresses from radio waves, uses a learned model generated by machine learning to differentiate between fixed and randomly changing MAC addresses, and estimates the number of persons based on this differentiation, considering additional attributes like time, facility type, and location.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If MAC addresses are collected from radio waves to estimate the number of persons, then the estimation process is simple, but the accuracy deteriorates when terminals have randomization function that changes MAC addresses frequently

Engineering Contradiction:

Improveestimation process complexityVSAvoidperson count accuracy

Core Design Contradiction:

Device complexityVSMeasurement precision

Solution Approach 1:

The patent segments MAC addresses into two categories: randomization MAC addresses (changed frequently) and non-randomization MAC addresses (remained constant). By separating the analysis of these two types, the system can accurately count persons carrying terminals with randomization function while maintaining simplicity in counting those without it.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent performs preliminary classification of MAC addresses by comparing historical data to identify which MAC addresses belong to terminals with randomization function. This preliminary action enables the system to differentiate between randomization and non-randomization MAC addresses before final counting, improving accuracy without significantly increasing overall complexity.

Inventive Principle:

Principle #10Preliminary action

2Productivity

If all MAC addresses are counted equally to estimate person number, then the calculation is straightforward, but the reliability deteriorates due to mixed randomization and fixed MAC addresses

Engineering Contradiction:

Improvecalculation efficiencyVSAvoiduser count reliability

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The patent segments the MAC address population into two distinct groups: randomization MAC addresses and non-randomization MAC addresses. This segmentation allows the system to apply different counting strategies to each group, maintaining calculation efficiency while improving the reliability of the final person count estimation.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent applies different quality standards to different parts of the data: randomization MAC addresses are analyzed using temporal patterns and historical comparisons, while non-randomization MAC addresses are counted directly. This local quality approach ensures each type is handled appropriately, enhancing overall reliability.

Inventive Principle:

Principle #3Local quality

3Reliability

If randomization function is enabled in terminals, then privacy protection is improved, but the accuracy of person estimation based on MAC addresses deteriorates

Engineering Contradiction:

Improveprivacy protectionVSAvoidMAC address-based estimation accuracy

Core Design Contradiction:

ReliabilityVSMeasurement precision

Solution Approach 1:

The patent converts the harmful effect of MAC address randomization (making counting difficult) into a beneficial feature by using the randomization pattern itself as an identifier. Terminals with randomization function exhibit predictable temporal patterns in their MAC address changes, which the system exploits to identify and count these terminals accurately while preserving user privacy.

Inventive Principle:

Principle #22Blessing in disguise (Convert harm into benefit)

Solution Approach 2:

The patent performs preliminary analysis of MAC address temporal patterns to identify terminals with randomization function before the actual counting process. This preliminary action enables the system to adapt its counting strategy based on the detected randomization behavior, maintaining accuracy despite privacy-protecting randomization.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS12198153B2Information processing device, information processing method, and system

Publication Date: 2025.01.14 UNERRY INC

AI summary

An information processing device includes an estimation unit estimating the number of persons in a facility based on a first number of MAC addresses counted based on first information about MAC addresses collected from radio waves transmitted by first terminals during a first time period, and a learned model, and a determination unit determining a third number of MAC addresses changing randomly, from the first number based on a first half of each of the MAC addresses. The learned model is generated by machine learning teacher data defining a relationship between a second number of MAC addresses counted based on second information about MAC addresses collected from radio waves transmitted from second terminals in the facility during a second time period, and the actual number of persons in the facility. The estimation unit estimates part of the number of the persons based on the third number and the learned model.