Quadrupedal Robot with Fluidic Muscles

For my Major-Qualifying-Project (MQP) at WPI, I, along with two other students constructed a highly modular, re-configurable skeleton made mostly with extruded aluminum bars. This extrusion allowed the team to completely reconfigure the geometry of the skeleton and experiment with the new characteristics in less than 5 minutes.

This skeletal platform is powered by a novel linear actuation system called "Hydro-muscles", which are latex tubes that exert forces using the principle of axial elongation under pressure due to radial constraint. This actuator is quite different from the more well-known McKibben muscle in that it is capable of power augmentation while still using an incompressible working fluid. This technology recently became the largest single intellectual property contract in the history of WPI as an institution. These muscles saw rigorous testing and validation while becoming integral components of many separate academic and research projects in Dr. Popovic’s lab, including this project..

A wide array of kinematic configurations where tested, with an equally wide spectrum of success. The most successful jump performed by the robot achieved a height of about 25 cm of ground clearance and 29 cm of forward motion.

In the end, the robot worked as planned, and became famous for being WPI's first quadruped robot capable of actual ambulatory motion on the ground, without support.

We were honored to be interviewed and have the story of our creation be told by the Boston Globe, Reuters, New York Times, and the WPI Alumni catalog.

Some links to these stories are at the bottom of this page.

This is a picture of my team. From the right, Dr. Marko Popovic (project adviser), Andres Leiro, Daniel Fitzgerald, and Thane Hunt (me!).

An artistic rendition of our robot dog from the WPI Daily Herd Campus Newspaper

We passed the mic around and improvised when Professor Popovic decided to give us a nice surprise by volunteering us to give a quick speech at the Cambridge Science Festival in 2015. It was fun, and best of all, our robot worked great during our dem

A very early prototype joint that I built when my team was only me. Here, the leg is fully constrained by springs rather than muscles, in order to allow us to experiment with the elastic dynamics of the quadruped without yet having implemented the a

We observed that a single muscle was not creating the highly energetic motion we expected, so I attached a second muscle to see what would happen. This worked acceptably, but I later determined that water flow (viscous drag) was the real problem to

After discovering that high-speed water doesn't resemble an ideal gas, we improved our plumbing to reduce the occurrence of large degree angles. I 3D printed some hydraulic muscle fittings with a less severe angle. It worked great, however great car

We ran a 5K as a team! I didn't know about it in advance, but I ran it in slacks to be a good sport. I kept assuring myself that my time was "at least better than our robot's".

prev / next

1 2 3 4 5 6 7 8 9 10 11 12 13

· · · · · · · · · · · · ·