Fall Detection Device Using Acceleration Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing fall detection devices face challenges in efficiently predicting falls, particularly when determining the start of a fall, due to complex processing requirements and interference from centrifugal forces during device rotation, leading to prolonged processing times and potential erroneous detections.
Innovation Solution
A fall detection device that includes a fall start monitoring unit, a low-gravity transition detection unit, and a falling state detection unit, which monitor changes in acceleration and differentiation values to predict falls without requiring simultaneous determination based on multiple threshold values, thereby reducing processing load and avoiding interference from centrifugal forces.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If both acceleration and differentiation value of acceleration are simultaneously used for determination, then fall detection accuracy is improved, but processing complexity increases and CPU load becomes large
Solution Approach 1:
The fall detection process is divided into two separate determination steps: first determining based on acceleration alone, then determining based on differentiation value alone. This segmentation allows each determination to use a single threshold value, reducing processing complexity while maintaining the ability to detect falls accurately through sequential evaluation
Solution Approach 2:
The system performs preliminary determination based on acceleration before performing the second determination based on differentiation value. This preliminary action allows the system to quickly filter out non-fall events using a simple acceleration threshold, reducing the processing load for more complex differentiation analysis
2Measurement precision
If two threshold values are used for determining fall state, then fall detection accuracy is improved, but processing time becomes prolonged
Solution Approach 1:
The use of two threshold values is segmented into two separate determination steps, where each step uses only one threshold value. The first determination uses an acceleration threshold, and the second uses a differentiation value threshold. This segmentation reduces processing time compared to simultaneously evaluating multiple thresholds, while still achieving accurate fall detection through the sequential process
3Adaptability or versatility
If acceleration sensor output is used for fall determination, then fall detection capability is achieved, but centrifugal force during rotation causes erroneous detection
Solution Approach 1:
The differentiation value acts as an intermediary that helps distinguish between acceleration changes caused by falling and those caused by rotation. By calculating the rate of change of acceleration and comparing it against a threshold, the system can identify the characteristic acceleration profile of a fall (rapid change followed by sustained low acceleration) versus the periodic acceleration of rotation, thereby improving detection reliability
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
The device effectively predicts falls, allowing for early measures to be taken, reduces processing time, and minimizes erroneous detections, enhancing the safety and responsiveness of portable electronic devices.
Implementation Method 1
a fall detection device for detecting a fall on the basis of an output signal of an acceleration sensor
Data Source
AI summary
By detecting that a differentiation value of the absolute value of acceleration is lower than a negative threshold value DAth1 in a negative direction from a first stage (S1: Stationary state) where the differentiation value of the absolute value of acceleration is approximately 0, it is considered that the stage is in a second stage (S2: Fall start state). When the absolute value of acceleration after that has become lower from a threshold value Ath1, it is considered that the fall detection device has entered a third stage (S3: Low gravity state). When the low gravity state has continued for a given time T3, it is considered that the stage is in a fourth stage (S4: Falling state), and a fall detection signal is output. Thus, the processing load when the determination is performed by software is reduced and the prediction of a fall can be performed.