Currently the single largest risk remains in the processing power requirement. We looked into Google Cartographer, which recommends 16GB RAM, but other projects have run Cartographer on embedded systems by using only 2D data like we are. We also plan to use ROS-provided SLAM libraries to do the fine-grain computations.

The only significant change made to our design was the inclusion of two more ultrasound sensors, facing forward and angled up/down. These are intended to address the concern of low obstacles that would otherwise go undetected by our Roomba sensors, but still bump into our mounted components. The inclusion of these sensors added little complexity or cost.

Schedule remains the same. Tomorrow we plan to finish the design report online and submit it. Next week during work sessions we plan to begin integration of the Pi, Roomba, and LIDAR.

All of our components have gotten off the ground. The Roomba is set up and we were able to control it via a USB tether to a computer. The Pi has been flashed, wireless set up, and ROS installed. The LIDAR had already been dry-run via a computer tether earlier. We have no more baseline unknowns about whether each component will work.