Train-A-Wear

A wearable workout monitor that provides real time feedback on the user's form while performing basic workouts to improve shape and exercise outcomes.

This uses a Raspberry Pi and three Inertial Measurement Units that detect position and orientation of the body segments they are placed on.

If you need help or want to brag about your sensor, find us on any social media or join our

Prerequisites

• Inertial Measurement Unit train-box
• Local WiFi network
• Android Smartphone

Server

• Linux machine with g++ compiler

Sensor software

• Arduino IDE
• ESP8266 Library
• LSM9DS1 SparkFun Library

Android app:

• An Android phone
  • Android SDK 27
  • Android Build tools v27.0.2
  • Android Support Repository

Installing

Server

To install the server run the following commands on a Linux machine:

cd ./Server
make
./tAw-server

And you would have the train-A-wear server running on port 31415. The port can be changed in the server file.

Microcontroller

Open the ino sketch from Sensor Software folder in Arduino IDE. Select settings for programming a generic ESP8266 board. Modify the following lines:

*ssid_1 = "X"
*pass_1 = "X"

with the values of your WiFi's SSID and password. Upload the sketch, reset the sensor and it will be up and running.

Phone application

Import the project from Phone App folder in Android Studio and run the gradle build. You can either generate an APK file or install it straight away on your phone.

Using the real time sensor system

Once everything is set up on the Raspberry Pi and the application is installed, the user can start using the features provided. The system requires a specific number and location for the inertial sensors, depending on the exercise type.

Setup

To obtain the correct sensor placement, consult the Wear it and go section inside each workout. This provides descriptions of the location and orientation, as well as images. The top of the sensor carcass displays two logos and two LEDs, which allows the user to determine the correct placement.

Quick start

Star the application -> Workout -> exercise of choice -> START

Obtaining feedback

Enter a workout using the steps quoted in Quick start. When the sensors are ready to go, the message "Ready to go!" is displayed. Then, the workout can be started anytime by pressing START, and stopped
The sensors send datagrams to the RPI, whose algorithms process it and send instructions to the app. These are displayed on the screen in each workout activity.

Additional features

To get help on network or sensor faults: start the app -> Help -> FAULTY SYSTEM For more detail on pre- and post-workout stretching: start the app -> Help -> RECOMMENDED STRETCHES Advice on healthy lifestyle is available: start the app -> Help -> HEALTH & LIFESTYLE

double Integration_algorithms::Roll(double gyro_y, double acc_x, double acc_z)
{
    double roll;
    double dt = 0.01;
    roll += gyro_y * dt;
    double acc_correction_y = atan2f(acc_z, acc_x) *180/M_PI;
    complementary_filter(&roll, acc_correction_y);
    return roll;
}

//rotation of the back (it does not stay stationary during the exercise)
//check for rotation along y axis
int Integration_algorithms::squat_straight_back(double gyro_y, double acc_x, double acc_z)
{
    double roll = Integration_algorithms::Roll(gyro_y, acc_x, acc_z);
    if (roll < 0.2 || roll > 0.2){
        return 1; //message "Bend during exercise!"
    } else{
        return 2; //message "Keep the nice posture"
    }
}

    public static String showFeedback (String instructionDef){
        switch (instructionDef){
            case "0":
                return("Ready to start");

            case "1":
                return ("Let's go!");

            case "2":
                return("Your back is bending, keep it straight while tensing your abs");

            case "3":
                return("Knees caving in: Keep your knees in line with your toes");

            case "4":
                return ("It looks like your form is asymmetrical when lowering in squat");

            default:
                return("You're doing great, keep going!");
        }
    }```