anymotion

Helping amateur athletes learn from their Heroes

Whether your an amateur looking to become like your Idol or a professional athlete, we at anymotion will help you become better at your chosen discipline.

What we learned?

How to work on the full stack from Azure over Application to Hardware.

How did we make it?

With determination and not enough sleep.

Components

• Posenet estimator for expert motion provided in flask api
• Arduino acceleration sensor code
• Flutter mobile app

Posenet

Posenet is implemented with detecron2 from facebook. Given a reference motion from a video, a target trajectory is generated.

Flask API

Hosted on Azure VM.

GET

• / Provides generated angle timeline of expert motion
• /video Provides video of pose estimation
• /angles Precomputed angles for demonstration

Arduino

Basic data pipeline using a kalman filter for state estimation.

Sensors

• Sen-MPU6050 Gyroscope & Accelerometer Module
• HC-05 Bluetooth Module

Flutter

State based prototype app for android and iOS. The app provides multiple exercise, animated in real time @60fps to compare.

Available exercises:

• Weightlifting
  • Bizep curl

Entry point for all connections regarding bluetooth and webservices. © 2019 - anymotion team - All Rights Reserved

Built With

• arduino
• python