PHONETECTION

How to dodge a war b/w life and death caused by a short 5 minutes delay to hospital after an accident?Phonetection!

Comment

Inspiration

For the past two decades there has been an alarming increase in the number of deaths due to road accidents which mostly go unattended or the assistance arrives late. We wanted to make a change in this by detecting accidents through a smartphone (Phone+Detection = Phonetection ;) )

What it does

The phone after a fixed position w.r.t car, basically gets your GPS and gyroscope data, after meticulated processing it detects an accident. After detecting, it asks the user whether it actually was one. If they reply with a no within 10 seconds, the app ignnores it. Otherwise, it detects it as an accident and finds the nearby hospital and police station and informs them with the precise location details and other personal details like name,blood group,etc. Also, an SMS is sent to their emergency contact.

How we built it

We used GPS and Gyroscope sensors which run as a service in the background of the app. Whenever an accident is detected, the phone wakes up and if it really were and accident, the coordinates of the smartphone and the personal details are sent to the server. Using Socket.io library, every organisation in the city(hospitals and police stations) are made aware instantly of the accident. Options are provided to takeup or pass the matter. The status of the victim in the meanwhile is constantly monitored and updated in the database as [attended, unattended and marked fit]. An SMS is dispatched to the emergency contact. And a search option is available to know the tentative location and status of the patient.

Challenges we ran into

We had to be very accurate in finding the conditions for an accident. Hence, the value for the speed and tilt-angle had to be precise. Since phones do not have sensors to calculate velocity and tilt-angles directly, we had to do indirect methods and available sensors to get the same. In the back end, sockets had to be monitored precisely, any abrupt disconnection could spoil the further alert systems and notifications. We had limited information to work the Mathematics on. Because the mobile phone doesn't give direct values of velocity and the angle made by it with the 3 axis, we had to come up with ideas to calculate them with the details provided by the phone hardware.

Accomplishments

By the end of the time, we finished almost every major features we thought of. Barring some basic UI challanges, we could successfully complete the technical algorithmic challanges that we faced, to link databases, to form links b/w data, to improve efficieny in the front-back end communications by equally distributing the amount of tasks and the things which we couldn't finish in time, we will take the next week to furnish the application.

What's next for PHONETECTION

We are planning to include more developers from our friend circles to make the product more precise and practical. We have to come up with better algorithms to scale up to a bigger user group. We are planning to use machine learning and AI to alert users to potential conditions which lead to future accidents. Notify people nearby with live data of crashes, so they can help the victim with early medication.

Share this project:

Updates