Data collection for machine learning occurs through the use of an accelerometer. The android system does not require permissions to use the accelerometer and gyroscope and consumes quite a bit of energy, which does not harm the level of charge of the device. At the moment, the application distinguishes between several types of physical activity – walking, running, cycling, sitting position and movement in transport.
The collection of data for subsequent analysis during research and experimentation
This application can greatly facilitate the collection of data for subsequent analysis during research and experimentation. The application allows you to collect and use data to build graphs and models for machine learning and study the operation of the accelerometer without tedious searches on the Internet. Setting up your experiments is now in your hands. There is no need to filter data taken from ready-made repositories if they do not meet the conditions of your research. You can build an experiment using a specific type of activity of the subject and limit the actions of the subject you do not need.
Freezio Sensor will increase the accuracy of your results
Freezio Sensor will increase the accuracy of your results and reduce the number of errors when calculating data. When using a mobile application, there is no need to purchase and configure additional body sensors – no bulky chest devices, which greatly speeds up and facilitates the process of building research.
Freezio Sensor instructions
Freezio sensor solves the problem of preparing for building a model and collecting data using the following technique:
1. When you click on the « Record » button, the application collects data from the gyroscope and accelerometer, as fast as they can provide them. On tested devices, the speed is from 200 to 1000 times per second.
2. After clicking on the « Save » button, the application transfers the accumulated data to firebase in the form of two files Accelerometr.txt and Gyroscope.txt (you can see them when saving the card). After sending, the card receives the status of Load and is displayed with a yellow check mark in the list.
3. backend server receives measurement data from firebase, performs calculation and adds calculation results from the card. After that, the card in the list is marked as green.
4. the mobile application can get the calculation result and build graphs.
Currently, the calculation is as follows:
1. Data is normalized – the accelerometer and gyroscope are placed in one csv file with a common axis t (time in ms). Resolution along the time axis – 0.01 sec.
2. Calculation of the angle of rotation of the telephone — integration of the gyro readings.
The plans add a calculation of the speed and height of the jump.
Freezio Sensor will help if necessary the analysis and collection of sports data. To achieve great sporting results in the modern world, all available means are used. Unfortunately, there is no single method for collecting data from an athlete for further processing. Using data from other athletes to analyze and improve performance on your team can lead to inaccurate calculations. Our application will allow you to analyze the specific sports activities of a particular athlete and achieve maximum accuracy of data analysis for the subsequent improvement of training.