Accelerometer Step Counting with Position-Adaptive Signal Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing pedometer technologies face challenges in achieving accurate step counting while minimizing false positives and maintaining low power consumption, often performing well in one area at the expense of another, and struggle to adapt to different body positions and walking styles.
Innovation Solution
The system employs a 3-axis accelerometer with signal processing circuitry that includes a low-pass filter, automatic position detection, event detection, classification, consistency checks, and noise rejection blocks, using statistical classifiers trained for specific body positions to generate accurate step counts and reduce false positives.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional pedometer algorithms are used to count steps, then the device can track step quantity, but false positives increase and accuracy decreases when the accelerometer is positioned at different body locations
Solution Approach 1:
The system dynamically adapts its step detection algorithm based on the detected body position. The processor identifies whether the accelerometer is worn on the wrist, hip, or another location, and automatically selects or adjusts the appropriate detection threshold and parameters. This dynamic adaptation allows the pedometer to maintain high accuracy across multiple wearing positions without requiring manual configuration.
Solution Approach 2:
The patent changes key detection parameters based on body position. Different acceleration thresholds, window sizes, and detection criteria are applied depending on whether the sensor is on the wrist, hip, or elsewhere. By modifying these parameters dynamically, the system resolves the contradiction between maintaining precision and adapting to versatility.
2Measurement precision
If complex signal processing algorithms are implemented to reduce false positives, then step counting accuracy improves, but power consumption increases
Solution Approach 1:
The signal processing is segmented into multiple stages: initial simple threshold-based detection, followed by more complex analysis only when needed. The system first applies lightweight filtering and basic event detection, then engages heavier processing algorithms only for ambiguous cases. This segmentation reduces overall power consumption while maintaining high false positive rejection rates.
Solution Approach 2:
The processor operates in periodic cycles, alternating between low-power monitoring mode and higher-power analysis mode. During normal operation, simple threshold detection runs continuously with minimal power. When potential step events are detected, the system periodically activates more sophisticated algorithms to confirm or reject the event, thereby reducing average power consumption while maintaining accuracy.
3Measurement precision
If multiple processing blocks are added to handle different body positions and reduce false positives, then step counting accuracy and false positive rejection improve, but device complexity increases
Solution Approach 1:
The processor is designed as a universal computing unit that can execute multiple different detection algorithms depending on the situation. Rather than having separate dedicated hardware circuits for each body position, a single programmable processor implements position detection, algorithm selection, and step counting functions. This multi-functionality reduces hardware complexity while maintaining the ability to handle various scenarios.
Solution Approach 2:
The system performs self-configuration by automatically detecting the body position and selecting appropriate parameters without user intervention. The processor analyzes the acceleration signal patterns to identify the wearing location, then autonomously adjusts its detection strategy. This self-service capability eliminates the need for complex manual configuration interfaces or multiple specialized processing paths.
Data Source
AI summary
An activity tracking device, such as a step-counting device includes an interface configured to receive one or more acceleration signals and signal processing circuitry. The signal processing circuitry generates an indication of condition of an accelerometer, such as a body position of the accelerometer, based on one or more accelerometer signals, generates an event signal, such as an event flag, based on one or more accelerometer signals and the indication of the condition of the accelerometer, and generates an activity signal, such as step flag based on the event signal, the indication of the condition of the accelerometer and one or more acceleration signals. The signal processing circuitry may generate a noise signal based on one or more acceleration signals and generate the activity signal based on the noise signal.