Close
Meet HAMR, the Harvard Ambulatory MicroRobot
Photo: Harvard Microrobotics Lab

My, my, my, my we just got hit, so hard 
By this robot, from Harvard
With sweet, autonomy 
In its mind to drive and four hype feet
It looks good, and we’ll bring you now
A super dope robot from Beantown
And it’s known as such:
This is a bot, you can’t touch.

Stop.

It’s HAMR time.

Here’s a bit of background on the Harvard Ambulatory MicroRobot, or HAMR (various incarnations of which Harvard has been working on for years), just to get you properly caught up:

One of the new and exciting things about HAMR in 2018 is the introduction of HAMR-F, which features onboard power and a first major step towards full autonomy. Previous versions of HAMR (we’ve written about some of them in the past) were mostly tethered and not particularly autonomous, which was okay, because Harvard was focusing on important stuff like manufacturing techniques and gait analysis. But to do practical things, you can’t very well run a very long electrical cord and stand there with a remote control all the time, can you? Or, you can, I guess, but nobody will take you very seriously if you do.

HAMR-F is 4.5 centimeters long and weighs just 2.8 grams, and it can zip along at up to 17.2 cm/s, or just under 4 body lengths per second. This is over 300 percent faster than the previous off-tether version of HAMR, and only a little bit slower than the tethered version, which is impressive. HAMR-F also features upgrades that “leverage advances in manufacturing, sensing, and energy storage to seamlessly integrate the electrical and mechanical components on HAMR.” An improved powertrain increases its robustness, and also gives the robot the ability to move much more dynamically, enabling aerial-phase motion, including everyone’s favorite gait, the pronk

HAMR-F features an improved powertrain that increases its robustness and also gives the robot the ability to move much more dynamically, enabling aerial-phase motion, including everyone’s favorite gait, the pronk

For off-tether operation, HAMR-F has an adorable little 8-mAh lithium-polymer battery. It can also be equipped with a slightly less adorable 25-mAh battery, giving it a runtime of 4.5 minutes. At top speed, its cost of transport (a dimensionless measurement of how efficient a thing is at moving itself) is about 84. To put this in context, the cost of transport of a typical cockroach is around 16. A human has a CoT of 0.2, but it’s more meaningful to use CoT to compare the efficiencies of robots that are similar in size. The researchers say that with some tweaks to HAMR-F’s speed and drive system, they may be able to get the robot’s CoT down to lower than the roach, which would be impressive.

At the moment, HAMR-F is controlled by a human. It’s a little bit wiggly, so the first thing to do in the interest of eventual autonomy was to get it to skitter in a straight line, and the addition of a MEMS gyro tamed HAMR-F to keep it within about 10 degrees of a commanded straight line. From here, the plan is to integrate additional sensors to improve control, and the researchers will also be exploring how to get HAMR-F to be useful in more challenging environments. Here’s a video highlighting the recent improvements:

HAMR-F is a project from Rob Wood’s Microrobotics Lab at Harvard, and the first author of this particular paper is Benjamin Goldberg. For more details on the most recent research, we spoke with Ben via email. 

IEEE Spectrum: How is HAMR-F able to move so much more quickly than its predecessors?

Benjamin Goldberg: There were two breakthroughs that allowed HAMR-F to be significantly faster than its untethered predecessors. First, HAMR-F incorporates a model-based redesign of the transmission from a prior publication. This redesign resulted in a significantly higher force output per unit mass. The second breakthrough is a compact and efficient tapped-inductor boost converter than can power eight independently driven, high-voltage, piezoelectric actuators. The combination of these two achievements makes HAMR-F an extremely fast mobile-microrobot-generalist that doesn’t need a tether or large power supply.

Can you describe the “large aerial phases and dynamic maneuvers” you are hoping to achieve with HAMR-F?

The goal is to make HAMR-F jump and run on its hind two legs. While HAMR-F can already run outside on rock and real-world surfaces, we want HAMR-F to be able to traverse larger obstacles and gaps. We’ve shown that these types of maneuvers are possible with previous tethered versions of HAMR, now we just need to translate this to be compatible with the added mass and microcontrollers onboard HAMR-F.

The paper says that the cost of transport for HAMR-F could end up lower than the CoT of a similarly-sized insect. How could you make that happen?

One of the reasons the cost of transport of HAMR-F could be lower than similarly sized insects is because of our use of specialized engineering materials. For example, the piezoelectric actuators we use have a higher power density and efficiency than biological muscle. This efficiency allows us to travel further on a single battery, or even utilize alternative power sources such as solar cells. The caveat here with a comparison to insects is that we aren’t at the point yet of having self-healing materials or hundreds of tactile-feedback sensing capabilities like insects do.

What are your plans for increasing HAMR-F’s autonomy?

We’ve shown simple controllers with HAMR-F, for example walking in a straight line, but we wish to expand on this by adding tactile feedback in the legs to detect obstacles or improve speed on surfaces with varying levels of friction. We are also exploring putting a camera onboard which would be able to record and transmit data and also be used for optic-flow based controllers.

Could you speculate about some potential future applications for HAMR-F?

One of the target applications for HAMR-F is to inspect and diagnose machinery or areas with limited accessibility— an engine or ductwork, for example. To realize these applications, we still need added sensors (e.g., a camera, gas composition), but these can be easily integrated due to the expandable circuit boards and payload carrying capacity of HAMR-F.

Yo, we outta here!

“Power and Control Autonomy for High-Speed Locomotion With an Insect-Scale Legged Robot,” by Benjamin Goldberg, Raphael Zufferey, Neel Doshi, Elizabeth Farrell Helbling, Griffin Whittredge, Mirko Kovac, and Robert J. Wood, has been submitted to IEEE Robotics and Automation Letters.

[ Benjamin Goldberg ] via [ Harvard Microbotics Laboratory ]

Special thanks to Elizabeth for assistance with the above adaptation of the lyrics of  “U Can’t Touch This” by MC Hammer.

The Conversation (0)

How the U.S. Army Is Turning Robots Into Team Players

Engineers battle the limits of deep learning for battlefield bots

11 min read
Robot with threads near a fallen branch

RoMan, the Army Research Laboratory's robotic manipulator, considers the best way to grasp and move a tree branch at the Adelphi Laboratory Center, in Maryland.

Evan Ackerman
LightGreen

This article is part of our special report on AI, “The Great AI Reckoning.

"I should probably not be standing this close," I think to myself, as the robot slowly approaches a large tree branch on the floor in front of me. It's not the size of the branch that makes me nervous—it's that the robot is operating autonomously, and that while I know what it's supposed to do, I'm not entirely sure what it will do. If everything works the way the roboticists at the U.S. Army Research Laboratory (ARL) in Adelphi, Md., expect, the robot will identify the branch, grasp it, and drag it out of the way. These folks know what they're doing, but I've spent enough time around robots that I take a small step backwards anyway.

The robot, named RoMan, for Robotic Manipulator, is about the size of a large lawn mower, with a tracked base that helps it handle most kinds of terrain. At the front, it has a squat torso equipped with cameras and depth sensors, as well as a pair of arms that were harvested from a prototype disaster-response robot originally developed at NASA's Jet Propulsion Laboratory for a DARPA robotics competition. RoMan's job today is roadway clearing, a multistep task that ARL wants the robot to complete as autonomously as possible. Instead of instructing the robot to grasp specific objects in specific ways and move them to specific places, the operators tell RoMan to "go clear a path." It's then up to the robot to make all the decisions necessary to achieve that objective.

Keep Reading ↓ Show less