I am currently learning ROS, as well as how to use a single-board computer like Beaglebone Black or Radxa Rock as the primary processing unit for a robot, with one or multiple microcontrollers (e.g. ATmega1284P) to offload processing requirements for sensor and actuator interfacing.
This project is my test bed for learning these techniques. I hope to eventually have an autonomous rover capable of navigating via GPS (pre-programmed waypoints) to complete various objectives.
I also want to use this project to learn image processing with OpenCV and to equip the robot with a medium-resolution camera. Ultimately, for the image processing capabilites, my goal is to enable it with facial recognition (i.e. the ability to recognize its builder) and to identify and distinguish between people.