Remember that 90's computer game where you had to guess where all the pesky little mines were? Well we want to bring that back, but in a much more real way. The presence of landmines and unexploded ordnance (UXO) is a global issue that affects many countries across the world. Here's some trivia for you: building a landmine costs only $3-$30, but disposing it can cost up to $1000.
So naturally we started thinking about how much it would actually cost to build an efficient landmine detection ecosystem using IoT that:
- minimizes human involvement,
- has a high detection rate,
- can be easily deployed on the field,
- uses lightweight equipment that won't trigger explosives,
- uses cheap sensors
...and it turns out it really wouldn't cost a lot.
What it does
Project Minesweeper uses lightweight programmable robots to 'sweep' areas prone to landmines. These bots collect real-time sensor data to infer possible threats using ML and alert the relevant authorities so they can properly neutralize them.
How I plan to build it
To test out the idea, we will use Anki's programmable robot: Vector and try to retrofit it with some cool new sensors that help us detect large suspicious objects and classify them. We will also attempt to build a Mission Control IoT dashboard that displays location data in real-time.
Challenges I ran into
Accomplishments that I'm proud of
What I learned
What's next for minesweeper.ai