Inspiration-
Medical emergencies are quite unpredictable in nature and they require immediate attention. However, it is not always possible for the person in need of help to communicate his needs. This is seen especially in cases of Seizure attacks , Tantrum attacks in case of Autistic children ,respiratory disorders, and cardiac arrests. Though the number of diseases that can be catered to are numerous, we decided to concentrate on Tantrum attacks in case of autistic children and Seizure attacks, since real time simulation of these attacks can be easily done for verification of the project through rapid hand movements in case of a attack. However, the system can be extended to other diseases and disorders.
How it works-
We have developed a wearable that will be worn by the autistic students or people suffering from Seizure attacks. Care was taken in choosing the wearable and the sensors so that they can be as non intrusive as possible. These sensors collect the accelerometer readings, galvanic skin resistance, atmospheric tempreture etc.
The raw data from these wearable sensors is sent to an edison board over bluetooth, where theraw data is unpacked and packed into JSON format and sent over to the cloud.
Cloud, which is hosted on openshift has the received data stored in a mongodb database. Analysis of the received data is done on the cloud. Since we are targeting Tantrum attacks in case of autistic children and Seizure attacks we make a statistical analysis (Standard deviation of a sliding window) of the Accelerometer readings and classify them into vigorous, walking movements and sitting.
Resulting Data from the cloud is sent over to an Webapplication where the results of analysis is shown. An alert in case of emergencies (vigorous hand movements) is sent through to the nearones of these people. An alert (buzzer)is also sounded to people around this person who can assist him. There is also a provision for analysing and tracking the data from multiple people at the same time on the web application.
Challenges we ran into-
Integration of more than one devices. Realtime streaming of data and analysis. Unpacking of raw sensor data and packing them into Json.
Accomplishments that we are proud of-
Successful detection of alert in realtime.
What we learned-
Connection of multiple sensors to the internet
What's next for Phoebus-
Development of more compact sensors Historical analysis of data collected Enhancement to the statistical analysis model Extending the system to other disorders
Log in or sign up for Devpost to join the conversation.