ProTech

Problem and Objective

Public shootings have become increasingly common over the past few years. In 2018 year alone, there have been 58 mass shootings and last year, there were a total of 367 mass shootings. In schools, many teachers try to help their kids run away and end up finding compromised hallways and routes. If we could give teachers and lost students alerts on where to go to escape and which routes and hallways were compromised. The students and staff could then find a way to safety.

Currently, schools use a lockdown system called A.L.I.C.E. It stands for Alert.Lockdown.Inform.Counter.Evacuate. The system is based on alerting others around you that a shooter is present. However, this current system puts the alerter in danger and does not tell those in hiding the best way to evacuate the building.

We wanted to create a system to help trapped students and teachers escape the grasp of school shooters.

Development

We started by outlining the parts of the system. We wanted to have a mobile app displaying the most accurate route to escape the shooter. We used a Flood Fill algorithm to detect the safest and shortest route to escape the building. We wanted to use a BOSE speaker to alert students trapped in corridors without phones on where the shooter is. To find the shooter we turned toward security cameras. Since security cameras are found almost everywhere and are inexpensive to implement, we wanted to use security cameras to find the shooter.

We used the whiteboard in our room to draw out a diagram of how our system would work.

We first wanted to collect all of our hardware. We talked to BOSE and pitched them our idea so we could get a speaker.

They thought our idea had a lot of potential and they loaned us a BOSE speaker to use in our project. Score!

We then split the work up and started working on the different parts of the system. The app consisted of a map that the user could touch to pinpoint where they were. They system then used data from a security camera ML script to find where the shooter was and used a flood fill algorithm to find the best possible route.

The server was built in Flask and takes input from a C++ program using ctypes. The flood fill algorithm is built in C++ since it is a lower lever language with high speed.

The ML program uses tensorflow and OpenCv to take pictures and runs a neural net to identify the shooter and sends an alert back to the flask server which communicates with the BOSE Speaker to alert people though sound.

The app then sends a get request to the server which then returns the array from the flood fill algorithm. The app then uses that to display the best possible route.

Demonstration

Due to wifi connection issues and no hotspot service in the event room, we weren't able to present our working solution. Here is a video of the system working:

Conclusion

We hope that our solution will be able to save thousands of live and help the future generation safe from unprecedented violence.