Automated Eating Detection via Accelerometer and CGM

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing methods for detecting eating episodes are cumbersome, prone to errors, and lack efficient, non-invasive, and cost-effective solutions for collecting training data to create personalized models for automated detection.

Innovation Solution

A system using a continuous glucose monitor (CGM) and an accelerometer to identify eating episodes by generating an individual model based on glucose and acceleration readings, allowing for the detection of eating episodes without relying on glucose data, and potentially incorporating additional sensors like PPG or heart rate data.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual food logging is used to monitor eating behavior, then individuals can track their food consumption, but the process is cumbersome and often skipped

Engineering Contradiction:

Improveeating detection accuracyVSAvoiduser convenience

Core Design Contradiction:

Measurement precisionVSEase of operation

Solution Approach 1:

The patent replaces manual food logging (mechanical/written recording system) with automated sensor-based detection systems including accelerometers, microphones, and visual sensors that automatically detect and record eating episodes without requiring user intervention

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system enables self-service by automatically monitoring and recording eating behavior through integrated sensors and processing units, allowing the device to serve itself in collecting dietary data without requiring active user participation

Inventive Principle:

Principle #25Self-service

2Loss of time

If end-of-day food logging is used to record consumption, then food intake can be captured, but timing accuracy is lost due to reliance on memory

Engineering Contradiction:

Improveeating episode timing precisionVSAvoidmemory-dependent accuracy

Core Design Contradiction:

Loss of timeVSReliability

Solution Approach 1:

The system implements continuous monitoring through accelerometers and other sensors that operate throughout the day, capturing eating episodes in real-time rather than relying on periodic end-of-day recall, thereby maintaining continuous data collection without gaps

Inventive Principle:

Principle #20Continuity of useful action

Solution Approach 2:

The system provides immediate feedback by detecting and recording eating episodes as they occur, with the processing unit analyzing sensor data in real-time to identify eating patterns and timestamp them accurately, eliminating delays associated with memory-based recall

Inventive Principle:

Principle #23Feedback

3Measurement precision

If camera-based visual sensors are used to detect food consumption, then eating episodes can be identified, but the system becomes complex and expensive

Engineering Contradiction:

Improvefood consumption detection capabilityVSAvoidsystem structure complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the detection system into multiple independent sensor modules (accelerometer, microphone, visual sensor) that can function independently or in combination, allowing the system to achieve accurate eating detection through multiple simpler components rather than one complex system

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system employs multi-functional sensors that can detect various types of eating behaviors (different foods, eating methods, contexts) using the same hardware platform, making the system universally applicable to diverse eating scenarios without requiring specialized equipment for each case

Inventive Principle:

Principle #6Universality (Multi-functionality)

4Measurement precision

If throat-mounted microphones are used to detect eating sounds, then eating episodes can be detected, but background noise interferes with accuracy

Engineering Contradiction:

Improveeating sound detection accuracyVSAvoidbackground noise interference

Core Design Contradiction:

Measurement precisionVSObject-affected harmful factors

Solution Approach 1:

The patent merges multiple detection modalities (accelerometer data, visual sensor data, microphone data) into an integrated analysis system that cross-validates eating episode detection across different sensor types, reducing reliance on any single sensor and thereby mitigating the impact of background noise on acoustic detection

Inventive Principle:

Principle #5Merging (Combining)

Data Source

PatentUS20230346306A1Methods and systems to detect eating

Publication Date: 2023.11.02 OPTUM LABS LLC

AI summary

Embodiments include methods and systems for automated eating detection. Systems comprise a continuous glucose monitor (CGM), accelerometer, and processing unit. During a first time period, the processing unit receives glucose readings and a first set of acceleration readings. The processing unit identifies an eating episode if the glucose readings satisfy one or more criteria. Using the identified eating episode and first set of acceleration readings, the processing unit generates an individual model that identifies eating episodes using acceleration readings without using glucose readings. During a second time period, a processing unit uses the individual model and a second set of acceleration readings to identify a second eating episode. Some embodiments may use additional sensor types (for example, PPG or heart rate) to identify eating episodes in the first or second time period, generate the individual model, or any combination thereof.