Robot Speed Control Using Face-Based Age and Gender Estimation

Solution Overview

Problem

Conventional methods for controlling the movement speed of service robots are inaccurate when users' bodies are partially obstructed, leading to unnatural speed adjustments and potential collisions, especially with children.

Innovation Solution

A device and method using a deep learning model to detect and track faces, estimate gender and age, and adjust movement speed based on a lookup table correlating these factors to optimize robot movement and user satisfaction.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If the robot uses conventional height detection methods to control movement speed, then it can identify users as children or adults, but it cannot accurately detect user height when body parts are covered by chairs or other objects

Engineering Contradiction:

Improveuser height detection accuracyVSAvoiddetection failure due to body coverage

Core Design Contradiction:

Measurement precisionVSObject-affected harmful factors

Solution Approach 1:

The patent introduces face detection as an intermediary method to indirectly determine user characteristics. Instead of directly measuring body height which is blocked by chairs, the system detects the user's face and estimates height from facial features, bypassing the obstruction problem entirely

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent replaces the mechanical/optical height measurement system with a deep learning-based facial recognition system. The controller uses AI algorithms to analyze facial images and estimate user characteristics, substituting physical measurement with computational analysis

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Ease of operation

If the robot adjusts movement speed based on partial body detection, then it can respond to user presence, but the speed adjustment becomes unnatural when body parts are covered and reappear

Engineering Contradiction:

Improverobot responsiveness to userVSAvoidmovement speed stability

Core Design Contradiction:

Ease of operationVSStability of the object's composition

Solution Approach 1:

The system continuously monitors facial images and provides feedback to the movement control. By tracking facial features across multiple frames and using deep learning estimation, the robot maintains stable speed adjustments even when users move or change position, preventing unnatural speed fluctuations

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The patent changes the detection parameter from body height (which varies with posture and obstruction) to facial characteristics (which remain relatively stable). The deep learning model estimates height, age, and gender from facial features, providing more consistent parameters for speed control

Inventive Principle:

Principle #35Parameter changes

3Device complexity

If the robot uses simple body detection methods, then the system remains simple, but it cannot accurately distinguish user characteristics when parts of the body are obscured

Engineering Contradiction:

Improvedetection system simplicityVSAvoiduser characteristic recognition accuracy

Core Design Contradiction:

Device complexityVSMeasurement precision

Solution Approach 1:

The patent changes the detection parameter from body height (which varies with posture and obstruction) to facial characteristics (which remain relatively stable). The deep learning model estimates height, age, and gender from facial features, providing more consistent parameters for speed control

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The patent replaces the mechanical/optical height measurement system with a deep learning-based facial recognition system. The controller uses AI algorithms to analyze facial images and estimate user characteristics, substituting physical measurement with computational analysis

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS12397436B2Device for controlling movement speed of robot and method therefor

Publication Date: 2025.08.26 HYUNDAI MOTOR CO LTD

AI summary

An embodiment device includes a memory configured to store a lookup table in which target speeds corresponding to genders and ages are recorded and a controller configured to detect a face from an image photographed by a camera provided in a robot, estimate a gender and an age corresponding to the face based on a deep learning model, search the lookup table for a target speed corresponding to the estimated gender and the estimated age, and determine the target speed as a movement speed of the robot.