Cycle Counting from Discontinuous Accelerometer Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing devices that monitor physiological parameters, such as the number of steps, face challenges in accurately counting periodic movements using discontinuous accelerometer data due to reduced visibility of signal periodicity, which affects battery life and user experience.
Innovation Solution
A system comprising an input unit for receiving discontinuous accelerometer data, a classifier to categorize movement into speed-based classes, a storage for periodicity values, and a movement calculation unit to determine the number of cycles based on classified movement classes and anthropometric data, allowing for precise counting of periodic movements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Use of energy by moving object
If discontinuous sampling is used to reduce battery power consumption, then energy efficiency is improved, but measurement precision deteriorates due to reduced visibility of signal periodicity
Solution Approach 1:
The system performs preliminary classification of movement types (walking, running, cycling) using discontinuous accelerometer samples before counting cycles. By pre-categorizing the movement pattern, the system can apply movement-specific parameters and models to accurately estimate total cycles even from limited data points, resolving the contradiction between sparse sampling and accurate measurement.
Solution Approach 2:
The patent introduces an intermediary classification step that bridges the gap between discontinuous samples and continuous cycle counting. The classifier acts as a mediator that infers movement characteristics from limited samples, enabling accurate cycle estimation without requiring continuous high-frequency data, thus maintaining measurement precision while enabling energy-efficient discontinuous sampling.
2Duration of action of stationary object
If discontinuous sampling is used to extend battery life, then duration of action is improved, but reliability deteriorates due to insufficient data for accurate threshold setting
Solution Approach 1:
The system changes parameters dynamically based on detected movement type. Different movement categories (walking, running, cycling) have different characteristic parameters including threshold values, sampling intervals, and counting models. By adapting parameters to the specific movement type rather than using fixed thresholds, the system achieves reliable cycle detection from discontinuous samples, extending battery life without sacrificing reliability.
Solution Approach 2:
The patent implements dynamic adaptation of detection parameters based on real-time classification of movement patterns. The system transitions from static threshold-based detection to dynamic parameter adjustment, where sampling frequency, threshold levels, and counting algorithms are optimized for the currently detected movement type, ensuring reliable operation across diverse activities while maximizing battery efficiency.
Data Source
AI summary
The present invention relates to device, system and method for counting the number of cycles of a periodic movement of a subject, e.g. for step counting of a person. To enable more precisely counting the number of cycles of a periodic movement of a subject based on discontinuous accelerometer data the proposed device comprises an input unit (21) for receiving discontinuous accelerometer data over time indicating movement of at least a body part of the subject, a classifier (22) for classifying the movement of the subject into one of several movement classes based on said discontinuous accelerometer data, each movement class being related to a different range of speed of a periodic movement, a storage (23) for storing a table including for each movement class a periodicity value per time unit, and a movement calculation unit (24) for calculating the number of cycles of the periodic movement of the subject per time unit, for a predetermined time or continuously over time by use of the periodicity of the movement class, into which the movement has been classified.