Betabot is a robot designed to work in conjunction with its simulator in order to put reinforcement and deep learning techniques into practice. It consists of a 3D printed body and legs, a Raspberry Pi Zero, gimbal motors & motor controllers, sensors, power boards, and a camera.
It runs Python, its own custom PID loops, OpenCV, and Betabot OS running on Raspbian.
Details
Betabot is designed to help you simulate an test reinforcement learning and computer vision techniques in the real world.
Take the 3D printed body, two motors, and six screws. Glue the small magnets onto the back of the motor axles. Place one motor on the side of the body. Rotate until screw holes line up, and screw three screws in half way.
Take one motor sensor board, and slide it into position under the three screws, with chip facing towards and sitting directly under the motor magnet. Screw three screws fully in to hold the motor and board in place. Repeat with the other motor.
Solder the power lead, via the on switch, 5v lead, and speed controllers onto the power board. Take the power board and and place it into the mounting square on the body, and screw it down with four spacing screws
Solder the 8 motor output header pins onto the controller board. Place the controller board on top of the power board, and screw it down.
Take the two speed controller boards and plug the motors into them. Place them down into the body.
Write the Betabot disk image to the SD memory card using https://github.com/tjacobs/betabot/blob/master/scripts/sd_restore, and pop it into the Raspberry Pi. Take the Raspberry Pi and place it on the body, with memory card at the outer edge. Screw down with four screws. Plug in the HDMI adapter, and OTG adapter.
Screw the legs onto the motors, and place the servos into the legs and feet, and screw together. Run the leads through the legs and into the controller.
Switch Betabot on. He should wake up after a minute and come to life!