Smart Dog Collar
Design and build a dog collar with embedded microcontroller to collect, store, and analyze movement data in order to detect seizure events with machine learning. This smart dog collar is able to send a push notification to a mobile app to alert an owner when a dog is experiencing a seizure.
Objectives
- Design an embedded device containing a microcontroller, accelerator, gyroscope, and battery. The design should seek to minimize footprint and power consumption. The chosen microcontroller (e.g., Arduino) must support blue-tooth and Wi-Fi and be able to run simple TensorFlow models.
- Build a simple mobile app (in Flutter) to test connectivity with the device and cloud-hosted database.
- Deploy existing machine learning model on the embedded device. The team will refine, improve, and test these models.
Motivations
Epilepsy affects 1 out every 130 dogs (0.75%). We desire to build a modern solution to identify these events to provide timely notifications to inform owners and to create detailed logs to be reviewed by veterinarians.
Qualifications
Minimum Qualifications:
- experience with microcontrollers
- dog lover
- experience with machine learning
- CS 492 (mobile) or CS 493 (cloud)
Details
Project Partner:
Patrick Donnelly
NDA/IPA:No Agreement Required
Number Groups:1
Project Status:Accepting Applicants
Website:https://soundbendor.org/assets/files/Symonds_NCUR21_poster.pdf
Video:
https://www.youtube.com/watch?v=Wk-hZgU_Mz0
Keywords:
Machine LearningSystem DesignEmbedded SystemsMicrocontrollerSensorsDogepilepsyseizureTensorFlow