Mechanical Design, Analysis, and Development
To call the UPenn mechanical curriculum intense would be an understatement. Between selected coursework and my own work in robotics research I gained exposure to a wide variety of disciplines. These included tools such as Finite Element Analysis, Computational Fluid Dynamics, and a plethora of Empirical Procedures. I continued to develop a knack for design that was discovered when I first started CAD work during the FIRST robotics competition in high school. All of these experiences show brilliantly in the three projects I took on with the Kod*lab robotics laboratory, two of which were designs I completed independantly as the only mechanical expert in an EE lab. Later on I took on several undergraduate mentees as they entered the lab.
Computational Fluid Dynamics
A great deal of my CFD experience came during our Senior Design Project on autonomous air-delivery of vital payloads using a small parafoil. The system was nick-named VATS or "Versatile Air Transport System." During the project we validated our payload, motor, prop, and parafoil selection using a 2D dynamic simulation. The simulation utilized 150 CFD analyses performed at varying angles of attack and free stream velocities. This information was compiled into a reasonable model of the parafoil's flight.
Finite Element Analysis
Many of the projects I took on in the Kod*lab involved improving upon the faults of previous systems, especially where fatigue and impact resistance were concerned. To that end, I performed a great many FEAs on complex assemblies of 20 components or more, often with simulated bolted constraints or composite materials. These simulations were used to justify improvement on problem areas from the past before the high cost and time investment of producing a prototype.
Empirical Testing Procedures
Often these preliminary analyses needed to be supplemented by hard empirical data. The experiments I crafted frequently utilized simple vision processing to measure metrics such as fluid flow rates or the modulus and hysteresis of a cantilevered composite beam. However I also was responsible for the empirical development of the Canid robot's (see below) first bounding behaviors. These first behaviors were hand tuned open loop gaits re-parameterized and tested over three hundred times since I began on the project.
For three years, beginning as an undergraduate in the summer of 2009, I worked as a research assistant in the Kod*lab; a subsidiary of the G.R.A.S.P. robotics laboratory at the University of Pennsylvania. While there, I was mentored by Kevin Galloway through the process of creating X-RHex, a complete mechanical redesign of an existing platform, and proceeded to independently execute the mechanical design and development of two more platforms. The X-RHex Light and Canid robots were designed in late 2010 then reviewed and prototyped in early 2011.
Computer Aided Design
Below are pictured final CAD renderings of the Canid, XRL, and X-RHex robots respectively. These CAD renderings were produced using the Photoview360 Solidworks Suite by Dassault Systems. The bulk of my experience in CAD and CAM is in Solidworks, FeatureCAM, SolidCAM, and to some extent Autodesk Inventor.
Manufacture and Development
The first physical prototype of each system is pictured below. Each system employs a variety of manufacturing processes in its construction. While X-RHex features complex components machined using a computer controlled 4-axis HAAS Mill, XRL features a frame composed of interlocking water-jet components. Canid utilizes a hybrid of both processes. Each robot also utilizes carbon and fiberglass composites for both compliance and strength. Both X-RHex and XRL were machined and constructed in-house by myself and others while Canid was built in collaboration with the Army Research Labs in Aberdeen Maryland.
The Canid robot was conceived with the intention of proving that quadrupedal legged animals derive benefit from the use of a flexible but compliant spine to improve the efficiency of their running/bounding. Working together with then Post-Doctoral Researcher and mentor Clark Haynes, we walked through the concept development process. After I had completed the initial CAD design, the project entered collaboration with the Army Research Labs and Jason Pusey. With their help we were able to build the first prototype and obtain preliminary behavioral results within a year.
Under the inspiration of Clark Haynes, I designed the first model of X-RHex Light in just two weeks. The platform was designed to do exactly what its name implies; to provide any and all functionality of the X-RHex robot at a reduced weight and with reduced manufacturing time and cost. To save weight, I envisioned a robot designed around an exoskeleton of carbon composite. The interior frame itself is extremely sparing. There are now four XRL robots in use around the Univeristy of Pennsylvania, Florida State University, and the Royal Veterinary College.
My original porject in the Kod*lab. X-RHex was designed to ameliorate shortcomings in previous RHex designs regarding weight, form factor, and durability. The new system has withstood intense and repeated abuse without fail, and delivers better internal sensory capacity, higher motor torques, and longer battery lives while still providing a relatively light 9kg base weight and payload room to spare on its dual picatinny rail mounts.
Laboratory on Legs:
An Architecture for Adjustable Morphology with Legged Robots
G. C. Haynes, Jason Pusey, Ryan Knopf, Aaron M. Johnson, and D. E. Koditschek
Proceedings of the SPIE Defense, Security, and Sensing Conference, Unmanned Systems Technology XIV (8387), April 2012.
X-RHex: A Highly Mobile Hexapedal Robot for Sensorimotor Tasks
Kevin C. Galloway, G. C. Haynes, B. Deniz Ilhan, Aaron M. Johnson, Ryan Knopf, Goran Lynch, Benjamin Plotnick, Mackenzie White, D. E. Koditschek
University of Pennsylvania Technical Report, 2010.