I am setting out to build the e-bike of kayaks! The kayak will be able to determine when the user is paddling and automatically turn on the electric motor to assist. Further, I'll train an ML model to detect if the user is trying to turn left or right. The ML model will use IMU (Inertial Measurement Unit) data from the oar trained against GPS data on the kayak to determine this.
The project is broken into four phases:
Phase 1:
Proof of concept showing e-kayak able to travel at walking pace for at least two hours
Phase 2:
Final form factor hardware / IMU integration
Phase 3:
Heading / fin GPS control
Phase 4:
ML Training and integration
Details
For full details on the build, check out my part one YouTube video!
Just finished my second video! Build out all the oar controller hardware and firmware on the breadboard. This video walks through all the technical details!
Next I'll be creating some PCBs to fit everything into the oar controller 3D printed housing!
Phase one of the project was the proof of concept. It showed I could build an e-kayak for under $1,000 that could travel at a walking pace with a range of at least 2 hours.
I ended up with a range of ~16 miles (4 hours) and a max speed of ~4.5 mph. I was super happy with the results and learned a lot through my first round of testing! Next step is to get my computer out of my lap and build up the oar hardware cluster!