Wearable IoT Behavior Recognition via IMU and Geomagnetic Fingerprinting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current human behavior recognition technologies face challenges in accurately detecting daily activities while ensuring privacy, as they often require multiple sensors, struggle with blind spots, and fail to distinguish between users or pets, and cannot determine specific behaviors like eating or using the toilet, due to reliance on cameras or wearable devices that only detect physiological signals.
Innovation Solution
A system combining an IoT network with a wearable device equipped with a nine-axis inertial measurement unit and altimeter, which uses finite state machines and fusion calculations to recognize user behavior by determining the specific functions of furniture in the environment, along with a physiological wristband for additional signal detection, and optional camera for behavior recording.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If cameras or multiple sensors are used to recognize human behavior, then behavior detection capability is improved, but system complexity and cost increase significantly
Solution Approach 1:
The patent extracts and removes the camera component from the behavior recognition system, relying instead on wearable devices with sensors to detect physiological signals and movement patterns. This eliminates the need for complex image processing while maintaining behavior detection capability through alternative means such as accelerometers, gyroscopes, and physiological signal sensors.
Solution Approach 2:
The wearable device serves multiple functions: it detects physiological signals (heart rate, blood pressure), tracks movement patterns, determines location, and recognizes behavior types. This multi-functional approach consolidates what would otherwise require multiple separate components into a single integrated device, reducing system complexity.
2Measurement precision
If cameras are used to capture motion images for behavior recognition, then behavior identification accuracy is improved, but privacy violations occur
Solution Approach 1:
The patent replaces the optical/mechanical camera-based system with a sensor-based detection system. Instead of capturing visual images that reveal personal information, the system uses accelerometers, gyroscopes, and physiological sensors to detect movement patterns and behavioral states, achieving behavior recognition without visual recording and thus protecting user privacy.
3Object-affected harmful factors
If wearable devices detect physiological signals to recognize activities, then privacy is protected, but behavior recognition accuracy decreases due to inability to distinguish specific actions
Solution Approach 1:
The patent adds temporal and contextual dimensions to physiological signal analysis. Instead of relying solely on static physiological data, the system analyzes time-series patterns of multiple sensors (accelerometer, gyroscope, physiological signals) and combines them with location data from IoT positioning. This multi-dimensional approach enables differentiation of specific behaviors such as eating versus reading, both of which may involve similar physiological states but distinct movement patterns and locations.
Solution Approach 2:
The system performs preliminary calibration and learning phases where it establishes baseline physiological patterns and movement characteristics for each user. By pre-processing and storing contextual information about user-specific behaviors, the system can more accurately recognize specific actions during actual use, improving precision without compromising privacy.
4Measurement precision
If indoor positioning systems use UWB for three-point positioning to locate users, then positioning accuracy is improved, but installation cost increases
Solution Approach 1:
The patent employs low-cost Bluetooth Low Energy (BLE) beacons instead of expensive UWB infrastructure for positioning. These inexpensive BLE beacons can be deployed widely throughout the environment to provide location data, achieving sufficient positioning accuracy for behavior recognition without the high installation and maintenance costs associated with UWB systems.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This system provides precise behavior recognition, including specific actions like urinating or brushing teeth, while ensuring privacy, by accurately identifying location, time, and habits, and can detect falls and prevent them, with the ability to monitor daily routines and physiological signals.
Implementation Method 1
The wearable device is equipped with a nine-axis inertial measurement unit (IMU) and an altimeter
Implementation Method 2
The wearable device is equipped with a nine-axis inertial measurement unit (IMU) and an altimeter
Implementation Method 3
The wearable device is equipped with a nine-axis inertial measurement unit (IMU) and an altimeter
Implementation Method 4
the packets provide the magnetic fingerprint, received signal strength indicator (RSSI) fingerprint, latitude and longitude, and user orientation of several key pieces of furniture
Implementation Method 5
the packets provide the magnetic fingerprint, received signal strength indicator (RSSI) fingerprint
Data Source
AI summary
A human behavior recognition system includes a wearable device and an IOT system; the wearable device has a built-in 9-axis inertial motion unit (IMU) and an altimeter, a wireless communication module, and a microprocessor; the user wears the wearable device which is fixedly attached to the chest to judge the user's actions including standing, sitting, lying, walking, running, roaming, falling, etc. The nodes or beacons of the IOT positioning system are installed in various areas of the living space. The packet information broadcast to the wearable device includes: the latitude and longitude of the preset installation location when the node or tag is installed, the area name where the installation location is located, and the name of each key furniture in the preset area, the corresponding geomagnetic fingerprint and RSSI fingerprint, corresponding latitude and longitude, and the user's orientation when using the key furniture; in this way, the wearable device can integrate the user's actions with the information of the broadcast packet, and directly calculate the user's behavior through the state machine in the wearable device to obtain the user's accurate behavior recognition. And further the system may add a physiological and biochemical signal detection bracelet for the user to synchronously detect the physiological and biochemical signals during the behavior, and then upload it to the server through the IOT system, then the server combines with user's habit, and conduct accurate behavioral analysis.