Agent Monitoring Devices With Learning-Based Selective Data Upload
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Monitoring devices for elderly, infants, and pets consume excessive power by continuously tracking locations and vital signs without considering the specific needs of the individual being monitored, leading to rapid battery drain and inefficient resource usage.
Innovation Solution
Implementing a learning model, such as LSTM, on the agent monitoring device to analyze sensor data and selectively upload data based on battery power levels and predefined conditions, using a local learning model to predict conditions and behaviors, and a cloud-based model for updates, thereby optimizing power consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If monitoring devices continuously track locations and vital signs, then monitoring accuracy and reliability are improved, but power consumption increases and battery life decreases
Solution Approach 1:
The system dynamically adjusts monitoring frequency and data transmission based on detected conditions. When the agent is in normal state, monitoring occurs at lower frequency to conserve power. When abnormal conditions are detected (e.g., falls, unusual activity patterns), the system increases monitoring intensity and transmission frequency, thus maintaining reliability while reducing average power consumption.
Solution Approach 2:
The system changes operational parameters such as sampling rate, transmission interval, and sensor activation based on battery power levels. When battery charge is high, the system operates at full monitoring capacity. When battery charge drops below thresholds, the system reduces monitoring frequency and prioritizes essential measurements, extending battery life while maintaining critical monitoring functions.
2Loss of information
If monitoring devices upload data continuously, then data availability and response time are improved, but power consumption and data transmission energy use increase
Solution Approach 1:
The system uses feedback from local learning model predictions to control data transmission. When the learning model predicts normal conditions with high confidence, the system suppresses data transmission to save energy. When predictions indicate abnormal conditions or low confidence, the system triggers immediate data upload, ensuring critical information is transmitted while minimizing routine transmission energy consumption.
Solution Approach 2:
Instead of continuously uploading all sensor data, the system selectively uploads only when necessary based on learned patterns. The local learning model processes data locally and triggers uploads only for significant events or when confidence thresholds are not met, reducing transmission frequency while maintaining data availability for important conditions.
3Device complexity
If monitoring devices use generic configurations for all agents, then device complexity and setup time are reduced, but monitoring precision and adaptability to individual needs decrease
Solution Approach 1:
The device performs self-customization by automatically learning each agent's baseline behavior patterns, activity cycles, and normal state characteristics through the local learning model. This eliminates the need for manual configuration while enabling personalized monitoring precision. The system adapts to individual agents autonomously, achieving both low complexity and high precision.
Solution Approach 2:
The system performs preliminary learning during an initial adaptation period to establish baseline behavior patterns for each agent. This preliminary action enables the device to later distinguish between normal variations and abnormal conditions specific to that agent, improving monitoring precision without requiring complex pre-configuration. The learning model is pre-loaded with general knowledge and fine-tunes individually for each agent.
Data Source
AI summary
A method implemented by an agent monitoring device, comprises obtaining, by a sensor of the agent monitoring device, sensor data over a period of time, the sensor data describing a characteristic of an agent associated with the agent monitoring device, determining output data for the sensor based on the sensor data using a learning model, determining a sensor condition for the sensor, determining that a power level of a battery of the agent monitoring device meets the pre-defined power level, determining whether the output data meets the threshold value of the sensor condition in response to the power level of the battery having reached the pre-defined power level, and uploading an indication of the output data to at least one of a cloud server or a representative device in response to the output data for the sensor having met the threshold value of the sensor condition.