LOW-COST ROBOTICS
During my freshman year of undergraduate studies at Rice University, I won a grant from the Oshman Engineering Design Kitchen (OEDK) to experiment with low-cost robotics. The aim of my project was to build a capable compute and sensing platform for less than $1000 and demonstrate its capabilities.
I built a system using a Nvidia TX-1 compute platform – a small but capable system-on-module with an 8-core ARM CPU and GPU. To accompany this, I used a combination of RGB and depth cameras to provide perception. To package everything, I designed and 3D printed a “head” unit that featured pan/tilt actuators for the cameras and kept the TX-1 housed.
In addition, for proof of concept, I built a 4 degree of freedom robot arm with off-the-shelf actuators and laser cut wood linkages. The arm was housed in an enclosure with mounting points for the head unit to which I attached a display. As my first foray into computational robotics, I implemented a basic motion planning and control pipeline using the Robot Operating System (ROS) and MoveIt!. The system used the perception suite to create an occupancy map and use state-of-the-art motion planning algorithms to move the robot’s end effector from point to point.
I entered my project in the annual Engineering Design Showcase in 2017 under the name “OpenArm”. My presentation focused on how, with some engineering, robotics is an accessible field and permits even the inexperienced to experiment with robots. This work won me the Excellence in Freshman Engineering Award and an additional grant for me to further explore this work.