Automated Self Driving Lifeboat

To prevent situations in which an unequipped and untrained person risks their own life trying to save others, I have decided to make a lifeboat, which can automatically search for a person, go to them, and be able to save them. I decided that this lifeboat has to be agile, stable and waterproof.

This can be controlled remotely, but at the same time is automatically able to detect a drowning person, and go to them without any human actions necessary. This is done through the use of Arduinos, and a relatively simple circuit inside of the boat itself.

Hardware:

1. 3D printer
2. Battery
3. Husky lens (https://www.dfrobot.com/product-1922.html)
4. Lifebuoy
5. Microbit Microcontroller
6. Propeller
7. Remote control Handle
8. Servo
9. small gas canister

Software:

1. Fusion 360
2. Mind+ coding software

After a lot of testing, I found that the best design would be that of a leaf-shaped bottom with a wing-like cross-section, as it seemed to have the maximum amount of lift with minimum air resistance.

I used Fusion 360 to build a model, firstly making the bottom two pontoons(3) by drawing out their general shape, then rotating around the axis. The two rectangles on top were then drawn out and added on. This was then copied and mirrored to make two separate pontoons on both sides to add to the boat

The lifeboat is split into two separate sections, both individually designed, with round pins on the top half which slot into the indents on the bottom half to allow a tight fit. You can see the two individual sections quite clearly in the first picture.

All of this is then sent to a 3D printer and printed.

The husky lens is attached to the front of the boat, as it contains an extremely useful facial recognition model. This allows the boat to identify people without having to do excess coding.

The propellers are attached to the back on both sides to allow steering, while the micro bit is added inside to keep it dry while allowing it to be wired to all parts of the boat. The battery should then be attached.

The servo is added and attached to the gas canister to release the lifebuoy if deemed necessary. This allows a quick deployment which is fail-proof, as only a simple mechanism is used.

The code is relatively simple (see above). This needs to be done in the Mind+ software, as it is compatible with the Husky lens. It uses information from the lens of the location of the person detected and turns using a difference in speed in the individual propellers.

As seen above, this boat is able to achieve its goal of speed, agility and automation very successfully! I've tested it several times since then, and am happy to say it is consistence in its performance.