Step Counter False Positive Rejection via Decision Tree Cross-Correlation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current portable electronic devices, such as smartphones and wearables, often register false positives when tracking steps due to movements like shaking or environmental vibrations, leading to inaccurate pedometer results.
Innovation Solution
A method that utilizes multi-axis accelerometer data to determine false positives by calculating statistical data and applying a decision tree with a cross-correlation mechanism, including a machine learning process, to differentiate between actual and false steps, and updates the decision tree based on user input and data collection during a configuration phase.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If simple step detection is used, then device complexity is reduced, but measurement precision deteriorates due to false positives from shaking or environmental vibrations
Solution Approach 1:
The step detection algorithm is divided into multiple independent decision stages (decision trees), each analyzing specific features of accelerometer data. The first decision tree evaluates basic step characteristics while the second decision tree performs cross-correlation analysis to eliminate false positives, allowing complex analysis to be broken into manageable segments that reduce overall computational complexity.
Solution Approach 2:
The system performs preliminary filtering and analysis of accelerometer data before final step confirmation. Decision trees pre-evaluate multiple parameters (magnitude, duration, frequency characteristics) to identify potential false positives early in the detection process, preventing unnecessary complex analysis and improving both speed and accuracy.
2Measurement precision
If complex false positive detection is implemented, then measurement precision is improved, but device complexity increases due to multiple decision trees and cross-correlation mechanisms
Solution Approach 1:
The complex detection algorithm is segmented into modular decision trees with distinct functions. Each decision tree handles specific aspects of false positive detection (magnitude thresholds, temporal patterns, cross-correlation), making the overall complex system more manageable and maintainable while achieving high measurement precision through coordinated analysis.
Solution Approach 2:
Decision trees serve as intermediary processing layers between raw accelerometer data and final step counting output. These intermediaries perform cross-correlation analysis and pattern matching to filter false positives, acting as a buffer that simplifies the interface between complex analysis requirements and simple step count output.
3Reliability
If statistical analysis and cross-correlation are performed, then false positive rejection is improved, but use of energy increases due to processing requirements
Solution Approach 1:
The system performs partial statistical analysis by selectively applying cross-correlation and decision tree evaluation only to accelerometer data points that meet preliminary threshold criteria. This partial action approach maintains high false positive rejection for suspicious data points while avoiding excessive energy consumption by skipping full analysis for clearly valid or clearly invalid steps.
Solution Approach 2:
The decision trees utilize configurable parameters and thresholds that can be adjusted based on activity type and environmental conditions. By dynamically changing analysis parameters, the system optimizes the balance between false positive rejection and energy consumption, applying more rigorous analysis only when necessary.
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 approach effectively reduces false positives, providing more accurate step counting by incrementing the step counter only for confirmed countable steps, thus enhancing the reliability of pedometer functions in electronic devices.
Implementation Method 1
receiving multi-axis accelerometer data representing a potential step taken by a user
Data Source
AI summary
Disclosed herein is a method including receiving multi-axis accelerometer data representing a potential step taken by a user of an electronic device. The method also includes determining whether the potential step represented by the multi-axis accelerometer data is a false. This determination is made by calculating statistical data from the multi-axis accelerometer data, and applying a decision tree to the statistical data to perform a cross correlation that determines whether the potential step is a false positive. If the potential step is not a false positive, a step detection process is performed to determine whether the potential step is a countable step and, if the potential step is found to be a countable step, a step counter is incremented.