A port of the ROS Arduino Bridge (https://github.com/hbrobotics/ros_arduino_bridge.git) using the Sparkfun Magician Chassis.

This project is a port of the ROS Arduino Bridge for the following hardware:

• Sparkfun Magician Chassis (https://www.sparkfun.com/products/retired/10825)
• Sparkfun Dual SEN12617 Encoder (https://www.sparkfun.com/products/12617)
• Raspberry Pi 2
• Alamode Arduino Raspberry Pi Shield (http://wyolum.com/projects/alamode/)
• Sparkfun Pan/Tilt Bracket (https://www.sparkfun.com/products/10335) w/ Servos
• Raspberry Pi Camera (TODO)
• Line sensor (TODO)

The hardware for this project essentially started with what is offerred on the Sparkfun website using a Raspberry Pi and the Dagu motor driver (DG-0813). Initially, the fine software provided by Dawn Robotics as seen here (https://bitbucket.org/DawnRobotics/raspberry_pi_camera_bot) was used and the general architecture is the same, i.e., Raspberry Pi to host the web server and receive commands from a web page and Arduino to receive sensor input and control the pan/tilt servos and drive motors.

The goal originally was to use ROS but the Raspberry Pi's at that time weren't powerful enough. When the Raspberry Pi 2 became available, using ROS was viable and this project began. Using the ROS Arduino Bridge as a baseline, extensions were added to support the Sparkfun hardware. The Dagu motor driver was replaced with the Alamode and Seeed Studio motor shields for a more compact design. The Arduino firmware was abstracted to create a single controller that supports a differential drive with encoders and PID control.

With ROS came the ability to modularize the project and take advantage of standard modules such as the ROS Bridge which provides a Javascript library for publishing and subscribing to topics. The web page has been reworked to use the ROS Javascript library.

Future additions to this project will include:

• using the ROS sensor/camera message to send images from the Raspberry Pi camera to the web page
• a line sensor for line following. Note: the ROS pub/sub model should make this an especially elegant solution
• using OpenCV to perform image processing such as facial recognition, object tracking etc.