Cypress' Maker of the Month: Will Bosworth | Cypress Semiconductor
Cypress' Maker of the Month: Will Bosworth
We are introducing a new section to the Cypress Core & Code blog: Maker of the Month! Through this series, we aim to recognize the unique projects created by the maker community. At Cypress, we are committed to helping our customers (international OEMs and makers alike) bring their innovations to reality with our easy-to-use software tools and programmable solutions. For more information about Cypress’s maker community involvement, check out our blog post on our presence at the recent Maker Faire Bay Area and our efforts to educate up-and-coming engineers via the Cypress University Alliance (CUA).
Read the Q&A with Cypress’s Maker of the Month, Will Bosworth. Will is a PhD candidate at MIT and has worked with his team to create the MIT Super Mini Cheetah using Cypress’s PSoC 5.
Please provide some background about yourself and your work at the MIT Newman Lab for Biomechanics and Human Rehabilitation and Biomimetic Robotics Lab.
I am a PhD candidate in the Mechanical Engineering Department at MIT, the Newman Lab is Professor Neville Hogan's research group and the Biomimetic Robotics Lab is Professor Sangbae Kim's research group. I began my PhD work with these groups as part of Sangbae Kim's MIT Cheetah project, which was funded by the DARPA M3 (Maximum Mobility and Manipulation) program.
Tell us about your Super Mini Cheetah project.
The MIT Super Mini Cheetah (SMC) is a small (9kg) inexpensive ($6,000) quadrupedal robot that can perform running, jumping and turning. The SMC robot builds on the design paradigm of the successful MIT Cheetah robot, which provides a robust approach to control leg force and impedance throughout a gait cycle.
The goal of this new robot was to create a low-cost, easy-to-use and easy-to-replicate platform that would enable a wide range of experiments. While the MIT Cheetah is an amazing legged robot, its size, strength and high cost can make the logistics of planning experiments very challenging. With a smaller, cheaper robot, we can take more risks: a single person can carry the robot to a new location or pick it up when it is misbehaving. And, if the SMC were to break, we could quickly and cheaply repair it. (I fully appreciate that $6k is not inexpensive for hobbyists—or even for my own personal finances!—but in this research space, the cost represents one or more orders of magnitude in hardware cost compared to many platforms).
To make an easily replicable platform, we emphasized the use of commercial off-the-shelf components and rapid manufacturing, and targeted a smaller robot size which could be transported and safely operated by a single person. The SMC enables low-cost, low-fear experimentation with legged robot hardware. The robot helps demonstrate that the cost of dynamic legged robots is dropping and will continue to do so.
In my PhD research, designing and building the SMC robot was merely "Step 1". The ability to run our machine outside over a variety of terrains (e.g., grass, dirt, gravel, concrete, etc.) has enabled us to identify and address some very interesting problems in legged locomotion research.
How did Cypress’s technology help you execute your project and what are the benefits you’ve seen with using Cypress’s technology?
The Super Mini Cheetah robot uses a PSoC 5 LP and I have been very impressed with it. When we were first building a single hopping leg in the summer of 2014, we began with the PSoC 4 development board, which is a nice size and conveniently priced. As we began to consider a four leg machine, we upgraded to the PSoC 5 for increased I/O and features. On the full SMC, we use a custom PCB based on the PSoC 5.
From my perspective, the PSoC platform really paid off for us when we moved from one prototype leg to a quadrupedal (four leg) robot. In many systems, quadrupling the number of I/O's might have resulted in significantly more software interruptions hiding out in our code (frightening!). But, the PSoC successfully abstracted away much of the I/O infrastructure for us. On our robot, the PSoC is doing the following I/O tasks: eight PWM's, eight quadrature encoders (actually, at some point we had to stop using the quadrature blocks, so we built our own edge-detector+counter blocks), four comparators, two UARTs (one for reading an IMU and one for Bluetooth communication). We're doing this I/O and some non-trivial onboard math and getting loop times of about 300 microseconds. It's really awesome!
A strength of the PSoC is that it was easy enough to use that we could get proof-of-concept code working quickly, yet powerful enough for us to consistently add new features to our robot/code.
Any additional comments you would like to add?
As an MIT student, the use of Cypress's products in the popular microcomputer class at MIT—Professor Steve Leeb's course 6.115—absolutely influenced my eventually discovering the PSoC. If you weren't aware: Prof. Leeb and his courses are absolutely amazing.
Along with my advisors—Prof. Sangbae Kim and Prof. Neville Hogan—and the general positive influences of their lab's members, the success of the Super Mini Cheetah project was heavily influenced by contributions from other MIT students who I had the good fortune of working with: Debbie Ajilo, Michael Farid, Jonas Whitney, Hans Susilo and Michael Chuah.
For questions or more information, you can reach out to Will Bosworth here.
Thanks again to Will for participating in our Maker of the Month program. If you would like to be considered for our next Maker of the Month profile, please email us and share a brief description of what project(s) you’ve made using Cypress’s technology.