Robots

My primary professional interest for the past 12 years has been designing and fabricating robots of all shapes and sizes, from 3-pound snake robot prototypes to 130-pound FIRST robots to 4,000-pound rideable hexapods. My life’s work, as I currently understand it, is to bring the fantastical creatures and experiences in my imagination to life through robotics; here are a few examples of that drive. Do you need any robots designed, mechanical engineering problems solved, CAD models made, or systems designed and integrated? Let me know – I’m happy to help.

Project Hexapod Logo

Stompy

Project Hexapod was founded by myself, James Whong, and Dan Cody, and is an effort to create an open-source, giant, rideable hydraulic hexapod robot called Stompy. Each of us had dreamed of one day building robots large enough to ride on, and we started working on the idea seriously in October of 2011. The effort took shape as a class at Artisan’s Asylum; over the course of 4 months, we taught a group of 15 students the ins and outs of working with robotic technology at large scales. Class exercises included creating robot control systems in simulation, designing and fabricating a half-scale prototype leg, designing a full-scale prototype leg, and starting development of the power unit and chassis. If you’d like to learn more about the class portion of robot development, check out my Robotics Intensive: Rideable Hexapod page. Ultimately, all of the development work led to the creation of a Kickstarter campaign to fund the actual robot:

We raised $97,817 with our Kickstarter campaign through the sale of swag (T-shirts, bumper stickers, wristbands, and more) and the offer of rides, drives, and sponsorships. The team then set to the task of developing our control systems on the full-scale prototype leg, while simultaneously working on the hydraulic power unit and mechanical design of the final robot.

My role on the team has largely been one of instructor and lead mechanical designer, responsible for the overall mechanical and hydraulic system design of the robot. Recently, I’ve also taken on the role of videographer; we’ve been posting video updates to our website to show our progress, and you can check them out at the Project Hexapod Youtube Channel.

We’re currently in the process of manufacturing the hydraulic powerplant, legs, and chassis of the final robot. We’ll soon switch gears to assembling, testing, and controlling the robot as a whole; we expect the robot to be done by late 2013.

ls3

LS3

LS3 (the Legged Squad Support System) is a four-legged, 1,200-pound, engine-driven hydraulic robot from Boston Dynamics that is intended to be the larger, more-capable, field-ready version of BigDog. The robot is intended to carry up to 400 pounds of equipment for a squad of soldiers and marines, with the capability of autonomously following them anywhere they might go on foot. LS3 is designed to cover at least 20 miles, or run for 24 hours, without stopping or refueling.

I had a number of roles over the course of this project. My initial involvement focused on overall body design and conceptual design of legs that would allow for extreme range of motion, squatting, and self-righting. Later, I focused on mechanical design for the legs and actuation systems. Finally, I was promoted into the role of Systems Integrator, assuming responsibility for developing the chassis of the robot, laying out all of the electrical, mechanical, hydraulic, and control components, and guiding the efforts of a large team of engineers and subcontractors to create one functional system.

pet-man

PETMAN

PETMAN is a tethered, hydraulic, anthropomorphic robot from Boston Dynamics that is designed to be the same size as, and as physically capable as, an average human male. It is intended to be used as a robotic mannequin to test chemical warfare clothing; it will eventually perform calisthenics in protective clothing while exposed to chemical warfare agents, in order to test the clothing’s capabilities.

In this program, I focused primarily on mechanical design of actuation systems, mechanical structures in PETMAN’s legs, and the mechanical design of the force-controlled, cable-tendon-driven intelligent safety harness that helps keep the robot balanced in case of system failure.

big-dog

BigDog

BigDog is a four-legged, 250-pound, engine-driven hydraulic quadruped from Boston Dynamics that remains one of the most nimble and maneuverable legged robots designed to-date. It’s capable of running at 4 miles per hour, climbing slopes up to 35 degrees, climbing muddy and snowy trails, and carrying up to 340 pounds over flat ground.

I joined Boston Dynamics late in this program, and as a result focused mostly on upgrades. I helped significantly improve part and system MTTF (mean time-to-failure) in the leg systems, sensors and the hydraulic powertrain, in particular.

robot-tuna

Olin Robot Tuna

My senior year capstone project at Olin College focused on building a 4-foot robotic tuna from scratch. I teamed up with Julia Buck, Paul Mandel, Erin Schumacher, Sarah Shiplett and Michael Taylor to develop this robot under the guidance of Professor David Barrett over the 2008-2009 school year for a SCOPE project sponsored by Boston Engineering. This effort was part of a Phase I SBTR (Small Business Technology Transfer) grant from the Office of Naval Research, held jointly by Olin College and Boston Engineering.

The robot is a hybrid of the design features of a regular submarine (i.e. dive planes, thruster-powered locomotion, and a rigid hull) combined with the flexible keel of a biological organism. This marriage produces a vehicle that can both move through the water quickly and turn on a dime, a set of traits not usually seen together in underwater vehicles of any kind. The tuna is used as a biological model because its natural swimming gait holds the front 2/3 of the fish’s body rigid, while the rear 1/3 moves; this allows the robot to utilize the front 2/3 of its body as a rigid, watertight hull, while the rear 1/3 is converted into a flooded flexible structure. The robot uses hydraulic actuators to move the flexible tail structure from side to side and electric motors for dive plane control. Two fully independent versions of this robot were developed over the course of 6 months; to accommodate such a tight development schedule, the robot was made almost entirely of fiberglass-reinforced rapid prototyping material.

Even though the tail structure only has one motor to move it from side to side, it still displayed the ability to swim by swishing its tail at a fixed frequency (without using its propeller to generate forward thrust). We believe that the robot could be engineered to swim at a relatively high cruising speed if its tail structure were tuned to an appropriate natural frequency, and if there was a sprung joint between the caudal fin and the rest of the tail. Unfortunately, the final robot was tested in the last days of my senior year, and we had no ability to test this hypothesis. Boston Engineering has since developed the project further into the GhostSwimmer and BIOSwimmer robots.

My role on the project included a 3-month stint as team leader and a full term as lead mechanical designer. The robot’s inherently flexible design continued the research into snake locomotion that I had been working on my sophomore year.

Snake Thumbnail

Olin Snake Robots

I took a class called Principles of Engineering in my sophomore year of college (2005-2006) at the Olin College of Engineering. The ultimate goal of the class was to introduce students to the engineering process, and over the course of the class create a product that was mechatronic in nature and controlled by a microcontroller. Me and a few other students (Matthew Aasted, Chris Dellin, Elizabeth Kneen, and Jon Tse) teamed up, approached Professor Gill Pratt for project inspiration, and decided to create snake robots in his newly-formed Biomimetic Robotics Lab at Olin.

We decided early on in the project to focus on a mechanical design unlike any other that we’d seen in a snake robot up until that point; instead of having a long series of joints, we decided to have a single elastic member down the entire length of the robot. Ribs attached to this ‘spine’, and hobby servomotors acted as muscles to push and pull the ribs against each other. This resulted in a robot that we call the ‘Serpentine Snake’ at the end of the fall semester of sophomore year, that moved forward by moving in a sinusoid from side to side and having passive wheels resist the urge to slip sideways (and instead, force the snake forward along its serpentine path).

Chris Dellin and I decided to continue the project after the fall semester as an undergraduate research project, and changed the design to allow for fully three dimensional movement – instead of only having muscles that allowed for side-to-side movement, we stiffened the actuators and gave the robot the ability to lift or lower its sections. This allowed the snake to achieve both a sidewinding gait and a rectilinear gait similar to an inchworm.

My role on this project was lead mechanical engineer and designer, responsible for the overall mechanical structure and actuation design of both robots. The robot parts were designed for lasercutter-based rapid prototyping.