Designing a Smart Dog Harness for Seizure Detection

Marji Symonds, Patrick Donnelly, Oregon State University-Cascades Computer Science Department, 1500 SW Chandler Ave. Bend, OR 97702

Epilepsy affects 1 out every 130 dogs (0.75%), a condition that affects both the well-being of the canine and owner. Although there are products on the market that measure vital signs in canines; there are no clinically researched products that have been shown to reliably identify when a dog is experiencing a seizure. We propose a smart-device to enable the future collection of data of seizure episodes. In this work, we present a system that combines a  microcontroller, an array of non-invasive sensors, and machine learning algorithms to attempt to detect seizure events in canines. We designed a wifi-enabled dog harness equipped with an accelerometer,  gyroscope,  GPS, and a microphone, coupled with a microcontroller programmed to analyze and log the sensor data. We then use this device to read data from rest, non-seizure activity, and simulated seizure episodes. Using this data, we train machine learning models to predict if a dog wearing this device is experiencing a seizure. Non-invasive reporting of seizure episodes in dogs will allow clinicians to fully track a patient’s seizure history and thus better inform the treatment of their epileptic patients. The development of this device will support the collection of data from epileptic dogs in future clinical trials. 

Additional Abstract Information

Presenter: Marji Symonds

Institution: Oregon State University

Type: Poster

Subject: Computer Science

Status: Approved

Time and Location

Session: Poster 5
Date/Time: Tue 12:30pm-1:30pm
Session Number: 4039