Digital IS
We Didn't Start the Fire!
Applying Swarm Intelligence to
Unmanned Aerial Vehicles for
Forest Fire Suppression
(A Broad Overview)
The amount of forest fires in the US are projected to increase by 50% by 2050
Forest fires cost the US government $683 million in 2008
How can technology help?
Idea:
Use Unmanned Aerial Vehicles (UAVs)
Use swarm intelligence to coordinate a group of UAVs
Like this, only without bombs
Use fire as
data input
Dropping chemical on the fire is a form of indirect communication between UAVs
UAVs contain the fire by following a path marked by chemicals
This way, UAVs only mark and trace the edges of the fire
Swarm Intelligence
Adapted from:
Results in complex behavior
We call this emergence
This is because complex behavior seems to "emerge"
from the interaction of individuals within the swarm
Interaction can be direct or indirect (through the environment)
Simulation
Uses Geographic Information Systems (GIS) data
Forest cover and elevation data
Swarm consists of 2-10 UAVs
Should be able to contain the fire using only the swarm
So what does the simulation look like?
Blue dots are UAVs
Gray squares are chemical deposits
Colored squares are fires
Fires are colored based on their temperature
Bluer colors = cooler
Whiter colors = hotter
Fires cannot cross gray squares
Color depends on tree species
Green areas are forests
Take another look:
See how the Pheromone Grid matches the Fire Grid?
It is an abstract representation of the fire
Each patch of fire has a pheromone
The greener the pheromone, the more attractive it is to UAVs
The darker the pheromone, the less attractive it is to UAVs
Blue pheromones are UAV pheromones
They are created when UAVs drop their chemical payload
This allows them to mark out a path and find the edges of the fire
That is a broad overview of my IS
Hope you enjoyed!
(actual data)
Set in fire-prone region of central Arizona
Problem
Big Ideas
Fire Grid
Pheromone Grid