We didn't set the fire, but we'll put it out!
The frequency and severity of wildfires has unfortunately increased in recent years. Just last year, California saw a wildfire that burned nearly two million acres and killed almost 100 people. A large part of this is due to the difficulty of containing wildfires once they start. Firefighters must directly risk their lives and equipment to enter danger zones in order to create firebreaks, or gaps in vegetation and combustible materials.
We believe an effective way to suppress wildfires is to contain them by stopping them from crossing highways. Quench is a prototype of an autonomous system that is intended to drive along a road and intelligently spray water at burning areas, in order to create and preserve fire breaks along existing infrastructure.
What it Does
Quench consists of a car that pulls a trailer. The trailer contains a camera, hose, water reservoir, and nozzle mounted on a pair of servos that support pan and tilt. The car can be controlled remotely or autonomously. As the car moves along, systems on the trailer continuously use computer vision to identify burning areas and spray water at areas that can best put out the fire.
How We Built It
The frame of the trailer is built from 80/20 aluminum, with key components mounted on.
The water suppression system was made by modifying a motorized water gun and mounting it to the frame. The nozzle of suppression system is connected to a pan/tilt servo system that allows for firing at any pitch and yaw. The water gun draws liquid from a refillable reservoir.
All computer vision processing and control is done locally on a Raspberry Pi computer, so there is no dependency on continuous internet access. We make use of the OpenCV library to identify the fire.
In order to pull the trailer, we modified an RC car. We reverse engineered the remote control of the RC car so that the motors could instead be driven from a Raspberry Pi via a motor shield.
Accomplishments We're Proud of
We're especially proud of our targeting system, and how it operates completely autonomously.
Quench is a small prototype, but we believe that this system can be scaled up to a practical car-sized vehicle that will help suppress wildfires and help prevent loss of life and property.
What we learned
We became familiar with the inner workings of OpenCV, as well as using a Raspberry Pi to drive motors. We also became pretty familiar with the inner workings of toy RC cars and water guns!
We hope Quench can scale up to a larger model that can be effective in fighting fires.