Motion Sensor Android Application

Background

Beat-PD data sets were collected using motion sensors on wearable devices such as smartwatch and mobile, and it was difficult to get an intuitive understanding of the data collection process, so we designed an application to collect data from mobile motion sensors to study the process of data generation.

Introduction

From the android developer documents, the device can obtain three different accelerations, including TYPE_ACCELEROMETER, TYPE_ACCELEROMETER_UNCALIBRATED, TYPE_LINEAR_ACCELERATION. The difference between them is if excludes gravity and bias compensation.

After a certain amount of data is accumulated, the data will be stored in a csv file and be exported to the cloud storage. By comparing our application generated data with the data in measurement, we found that measurement data did not exclude the gravity because the mean of the time-series is not 0.After using the application to simulate the various users might some daily activity and symptoms in patients with Parkinson’s disease(swing arm, walking, etc).The trend in the data may be interpreted as is a large movement of daily activities. We cannot assume that these results are exactly the same as real data collection methods, but they should be similar.

User Interface

Generated Data

Motion Sensor Application Download

https://github.com/GoldExperience/MotionSensor/releases/tag/v1.0