One of the priorities to improve the first prototype was to get rid of the need to connect Babot to a computer to make it work. Since the position of the ball is detected by a camera, it requires something powerful enough to do computer vision. Using a Raspberry Pi was the perfect solution, it’s relatively simple to use and makes Babot’s code super easy to modify by anyone. However a Raspberry Pi is not as powerful as a traditional computer so I had to rewrite all the Python code to optimize it in order to run the computer vision code in real time. To give you an idea Babot runs at around 40 FPS with a camera set to a very low resolution to reduce computing time.
The main ports of the Raspberry Pi are fully accessible from the outside of Babot. The Ethernet port is very useful if you want to reprogram Babot. The other USB ports can be used as you wish. You could for example plug in a computer mouse or a joystick to control the position of the ball on the plate or you could plug several Babots together and make them communicate with each other via USB.