Hi everyone,
New here, so I guess a big Hello is a good idea 
I'm a young French computer science folk. I loved doing mecanical stuff with my Mecano when I was around 10. It's been a long time since I craft something. After maths and computer science studies, I have been studying the whole pack of machine learning field for more than one year now. Recently, I've been involved in a reinforcement learning (RL) course. Amazing teacher, amazing ideas behind it.
I wanna build a robot capable of moving around without hitting all and each obstacles around him.
I want it to be able to get instructions over Wi-Fi. The idea behind that is to reproduce the enviromnent on a RL simulation in order to train my agent(the robot) to avoid obstacles (the context).
My solution would be to use a Wi-Fi camera (or a smartphone) to send some pictures from the robot environment to a network port where resides my RL based decision tool. The microcontroller will listen to that port and take instructions from it.
I came to that roadmap in order to achieve that goal :
Step 1 - Materials
- 1 Wi-Fi microcontroller board
- 1 Wi-Fi camera
- 2 motors
- 1 energy supply
- some connectors I guess to connect motors to the microcontroller, to the energy supply, ...
- mecanical parts (chassis, wheels, ...)
Step 2 - Assembly all the stuff
Step 3 - Send instructions from Wi-Fi to the robot
Step 4 - Develop RL environment
Step 5 - Enjoy good stuff
I have ton of interrogations about it but more important ones to start are :
- What do you guys think of the process overall ? Is it feasible ? Is there a better way to achieve it ?
- Which microcontroller board should I buy in order to be listening to a remote server ? (I saw ARDUINO UNO WiFi REV2 — Arduino Official Store)
- Which Wi-Fi camera and motors will be compatible with that microcontroller board ? (No idea there)