Yingnan Ma

Pre-impact and Impact Detection of Falls Using Built-In Tri-accelerometer of Smartphone

Falls in elderlies are a major health and economic problem. Research on falls in elderly people has the great social significance under the population aging. Previous smartphone-based fall detection systems have not both fall detection and fall prevention, and the feasibility has not been fully examined. In this paper, we propose a smartphone-based fall detection system using a threshold-based algorithm to distinguish between Activities of Daily Living (ADL) and falls in real time. The smartphone with built-in tri-accelerometer is used for detecting early-warning of fall based on pre-impact phase and post-fall based on impact phase. Eight healthy Asian adult subjects who wear phone at waist were arranged to perform three kinds of daily living activities and three kinds of fall activities. By comparative analysis of threshold levels for acceleration, in order to get the best sensitivity and specificity, acceleration thresholds were determined for early pre-impact alarm (4.5-5m/s2) and post-fall detection (21-28 m/s2) under experimental conditions.

Feature Selection and Predictors of Falls with Foot Force Sensors Using KNN-Based Algorithms.

The aging process may lead to the degradation of lower extremity function in the elderly population, which can restrict their daily quality of life and gradually increase the fall risk. We aimed to determine whether objective measures of physical function could predict subsequent falls. Ground reaction force (GRF) data, which was quantified by sample entropy, was collected by foot force sensors. Thirty eight subjects (23 fallers and 15 non-fallers) participated in functional movement tests, including walking and sit-to-stand (STS). A feature selection algorithm was used to select relevant features to classify the elderly into two groups: at risk and not at risk of falling down, for three KNN-based classifiers: local mean-based k-nearest neighbor (LMKNN), pseudo nearest neighbor (PNN), local mean pseudo nearest neighbor (LMPNN) classification. We compared classification performances, and achieved the best results with LMPNN, with sensitivity, specificity and accuracy all 100%. Moreover, a subset of GRFs was significantly different between the two groups via Wilcoxon rank sum test, which is compatible with the classification results. This method could potentially be used by non-experts to monitor balance and the risk of falling down in the elderly population.

A wearable pre-impact fall early warning and protection system based on MEMS inertial sensor and GPRS communication

Fall is one of great threats affecting people with old age. Aimed at the falling issue of aged, the paper explored a pre-impact fall early warning and protection system. This system consists of an early fall alarm, protection airbags, a remote monitoring platform and a guardians cellphone app. The early fall alarm and airbags are integrated in a belt, convenient for wearing and hip-protection. The inner of early fall alarm has a MEMS sensor, which collects 3-axis accelerated velocity and 3-axis angular velocity. A fall detection algorithm is applied to recognize falls from activities of daily living (ADL). When there is a dangerous movement approaching fall, the early fall alarm will warn the aged to stop the movement. When the fall happens, the early fall alarm will trigger the airbag system, then the airbags in the belt will inflate as soon as possible to reduce the damage to the aged. In addition, the early fall alarm will ring and send message to the guardians cellphone for help. Meanwhile, the kinematics of the human body during falling time will be stored in TF card and sent to remote monitoring platform for storage. Then the monitoring platform can show the fall location where the fall incident happens in the electronic map. In order to test the reliability of this early fall alarm and protection system, a series of experiments have been designed. The results show that this system can be relatively accurate to detect falls, accomplishing functions including early fall warning and alarming, airbag inflation, statics transferring and storage, real-time location, which has significant benefit for reducing the direct damage and shortening the aiding time.

Accurate Fall Detection Algorithm Based on SBPSO-SVM Classifier

For the purpose of improving the medical care which aims at the elderly and the chronic patients who are prone to falls, this paper makes use of Standard Binary Particle Swarm Optimization(SBPSO) to search for the combination of best feature subset and parameters (C, g), which can be used to train the SVM(Support Vector Machine). Experiments results show that the proposed method can get higher accuracy (about 99%) compared with non-optimized SVM, k-NN (k Nearest Neighbors) and threshold-based method when dealing with the classification of ADL (Activities in Daily Life) and abnormal falls.

Relationship between dynamical characteristics of sit-to-walk motion and physical functions of elderly humans

Sit-to-walk (STW) motion is essential for daily activities. Falls frequently occur, when there is impaired ability to perform STW movements. This study investigated the relationships between the dynamical characteristics of STW motion and physical functions of elderly people. 128 elderly (51 males and 77 females, above 65 years) participated in this study. Participants were instructed to perform STW motion at comfortable state and classified into four groups (normal, mild, moderate and severe group) based on physical function, which evaluated with functional reach test and dynamic gait test. The results showed that some relationships were confirmed between the sample entropy characteristics of STW motion. Moreover, a subset of variables was significant different among four groups via Kruskal-Wallis test, which could be potentially used for developing an objective and simple methods to assess balance capacity and fall risk level of elderly people.Sit-to-walk (STW) motion is essential for daily activities. Falls frequently occur, when there is impaired ability to perform STW movements. This study investigated the relationships between the dynamical characteristics of STW motion and physical functions of elderly people. 128 elderly (51 males and 77 females, above 65 years) participated in this study. Participants were instructed to perform STW motion at comfortable state and classified into four groups (normal, mild, moderate and severe group) based on physical function, which evaluated with functional reach test and dynamic gait test. The results showed that some relationships were confirmed between the sample entropy characteristics of STW motion. Moreover, a subset of variables was significant different among four groups via Kruskal-Wallis test, which could be potentially used for developing an objective and simple methods to assess balance capacity and fall risk level of elderly people.

Fall risk factors analysis based on sample entropy of plantar kinematic signal during stance phase

Falls are a multi-causal phenomenon with a complex interaction. The aim of our research is to study the effect of multiple variables for potential risk of falls and construct an elderly fall risk assessment model based on demographics data and gait characteristics. A total of 101 subjects, whom belong to Malianwa Street, aged above 50 years old and participated in questionnaire survey. Participants were classified into three groups (high, medium and low risk group) according to the score of elderly fall risk assessment scale. In addition, the data of ground reaction force (GRF) and ground reaction moment (GRM) was record when they performed walking at comfortable state. The demographic variables, sample entropy of GRF and GRM, and impulse difference of bilateral foot were considered as potential explanatory variables of risk assessment model. Firstly, we investigated whether different groups could present difference in every variable. Statistical differences were found for the following variables: age (p=2.28e-05); impulse difference (p=0.02036); sample entropy of GRF in vertical direction (p=0.0144); sample entropy of GRM in anterior-posterior direction (p=0.0387). Finally, the multiple regression analysis results indicated that age, impulse difference and sample entropy of resultant GRM could identify individuals who had different levels of fall risk. Therefore, those results could potentially be useful in the fall risk assessment and monitor the state of physical function in elderly population.Falls are a multi-causal phenomenon with a complex interaction. The aim of our research is to study the effect of multiple variables for potential risk of falls and construct an elderly fall risk assessment model based on demographics data and gait characteristics. A total of 101 subjects, whom belong to Malianwa Street, aged above 50 years old and participated in questionnaire survey. Participants were classified into three groups (high, medium and low risk group) according to the score of elderly fall risk assessment scale. In addition, the data of ground reaction force (GRF) and ground reaction moment (GRM) was record when they performed walking at comfortable state. The demographic variables, sample entropy of GRF and GRM, and impulse difference of bilateral foot were considered as potential explanatory variables of risk assessment model. Firstly, we investigated whether different groups could present difference in every variable. Statistical differences were found for the following variables: age (p=2.28e-05); impulse difference (p=0.02036); sample entropy of GRF in vertical direction (p=0.0144); sample entropy of GRM in anterior-posterior direction (p=0.0387). Finally, the multiple regression analysis results indicated that age, impulse difference and sample entropy of resultant GRM could identify individuals who had different levels of fall risk. Therefore, those results could potentially be useful in the fall risk assessment and monitor the state of physical function in elderly population.

A quaternion-based Attitude Estimate System Based on a Low Power Consumption Inertial Measurement Unit

Accurate and real-time tracking of the orientation or attitude of rigid bodies has traditional applications in robotics, aerospace, underwater vehicles, human body motion capture, etc. Towards human body motion capture, especially wearable devices, the use of a longer time has always been a challenge for several weeks or several months continuously, so a low-cost chip and a low computational cost algorithm are necessary .The paper presented a quaternion-based algorithm that integrated the sensor output with the Kalman filtering algorithm, and a low power consumption Inertial Measurement Unit (IMU) for the attitude estimation. The low power consumption IMU with an inner Digital Motion Processor(DMP) from InvenSense Inc. called MPU9150, which contains triaxial accelerometers, triaxial gyroscopes, triaxial magnetometers and inner DMP. Firstly, we got attitude quaternion from DMP, and used the factored quaternion algorithm (FQA) to calculate course angle quaternion component. Then the Kalman Filtering algorithm was used to mix them together to acquire the accurate and good real-time performance attitude .The experimental results showed that Kalman filtering algorithm to mix DMP output and magnetometers data have better performance than gradient descent algorithm and complementary filter algorithm even in static performance and dynamic performance and power consumption.

Fall Detection with the Optimal Feature Vectors Based on Support Vector Machine

Falls have caused extensive interest of the researchers for it becomes the second largest accidental injury to death in the world. And there are lots of approaches to fall detection at present. However, on account for the complexity of this problem, a preferable effective method for fall detection hasn’t been present so far. This paper adopts a relatively high-predicted and stable SVM classifier to predict falls. 10 healthy young subjects participated in this study based on the Xsens MVN Biomech system. With the extraction of feature vectors, as well as the exploration of the best position, it found that the waist would be the best to measure body’s motion, and the simple accelerometer can offer the preferable features for the classifier to determinate the falls well. Meanwhile it can get a high accuracy up to 96% by setting an optimal C and g with five-fold cross-validation testing.