Inspiration
The first thing that comes to our minds when we think of Hajj-related issues is the incident of September 2015 where more than 700 pilgrims were dead (رحمة الله عليهم). This was the main inspiration that motivated us to think of a solution that treats Crowd Management issues.
What it does
It detects the overcrowded areas and acts accordingly by guiding people to less crowded and safer places using drones and applying machine learning algorithms. The movement of pilgrims will be monitored by surveillance cameras attached to the drones allowing the detection of crowded areas using machine learning that recognizes the number of pilgrims per square meter and decides whether they're in danger or not. If the spot is judged overcrowded, a notification is sent to the admin who can access a live streaming of what's happening around the drone by changing the camera angle, he can then generate a greenlight projector (attached to the drone ) to show the pilgrims the less crowded way to go by,if the surrounding areas are also overcrowded a STOP signal is generated ( using an orange light ). The concerned pilgrims will stop until there's a freeway again which will also be announced by the green light projector generated by the admin. The admin has access to a heat map generated by the data collected by the different drones and containing the positions of the different drones. By clicking on a drone the corresponding information is displayed such us (live video streaming and statistics)
How we built it
We simulated the treatment that should be done by the drone embedded system by a mobile app (Android app) that takes photos and treats them using machine learning (human face detecting and counting the number of people in the photo) and we developed the software used by the admin ( using MDC (Mobile Device Components ) and then the framework Materialize as for backend we used Node.js)
Challenges we ran into
the limited time (which meant we couldn't train a new model and had to work with an existing one) Nonavailability of the hardware
Accomplishments that we're proud of
Being able to develop the machine learning face detecting feature in Android. Being able to work with a new framework (Materialize and MDC ) and adapt to them in less than 24 hours. Being able to work together as a team by exploiting the potential of each member Being able to deliver a product at the end of the hackathon
What we learned
Discovering new APIs and exploiting their documentation Stress management Teamwork Time management Effective ways of pitching and public speaking
What's next for A-021-Manfadh
Manfadh has so many potential in the future, with the evolution of research related to drones and machine learning. We can add a data analysis feature that provides detailed data and statistics about all the pilgrims We can also use drones to deliver first aid in crowded areas where help is needed and is impossible to be reached by human agents. The ability to recognize faces and identifying them in order to look for pilgrims declared lost.
Log in or sign up for Devpost to join the conversation.