Tag Archives: mechanical

#436044 Want a Really Hard Machine Learning ...

What’s the world’s hardest machine learning problem? Autonomous vehicles? Robots that can walk? Cancer detection?

Nope, says Julian Sanchez. It’s agriculture.

Sanchez might be a little biased. He is the director of precision agriculture for John Deere, and is in charge of adding intelligence to traditional farm vehicles. But he does have a little perspective, having spent time working on software for both medical devices and air traffic control systems.

I met with Sanchez and Alexey Rostapshov, head of digital innovation at John Deere Labs, at the organization’s San Francisco offices last month. Labs launched in 2017 to take advantage of the area’s tech expertise, both to apply machine learning to in-house agricultural problems and to work with partners to build technologies that play nicely with Deere’s big green machines. Deere’s neighbors in San Francisco’s tech-heavy South of Market are LinkedIn, Salesforce, and Planet Labs, which puts it in a good position for recruiting.

“We’ve literally had folks knock on the door and say, ‘What are you doing here?’” says Rostapshov, and some return to drop off resumes.

Here’s why Sanchez believes agriculture is such a big challenge for artificial intelligence.

“It’s not just about driving tractors around,” he says, although autonomous driving technologies are part of the mix. (John Deere is doing a lot of work with precision GPS to improve autonomous driving, for example, and allow tractors to plan their own routes around fields.)

But more complex than the driving problem, says Sanchez, are the classification problems.

Corn: A Classic Classification Problem

Photo: Tekla Perry

One key effort, Sanchez says, are AI systems “that allow me to tell whether grain being harvested is good quality or low quality and to make automatic adjustment systems for the harvester.” The company is already selling an early version of this image analysis technology. But the many differences between grain types, and grains grown under different conditions, make this task a tough one for machine learning.

“Take corn,” Sanchez says. “Let’s say we are building a deep learning algorithm to detect this corn. And we take lots of pictures of kernels to give it. Say we pick those kernels in central Illinois. But, one mile over, the farmer planted a slightly different hybrid which has slightly different coloration of yellow. Meanwhile, this other farm harvested three days later in a field five miles away; it’s the same hybrid, but it also looks different.

“It’s an overwhelming classification challenge, and that’s just for corn. But you are not only doing it for corn, you have to add 20 more varieties of grain to the mix; and some, like canola, are almost microscopic.”

Even the ground conditions vary dramatically—far more than road conditions, Sanchez points out.

“Let’s say we are building a deep learning algorithm to detect how much residue is left on the soil after a harvest, including stubble and some chaff. Let’s drive 2,000 acres of fields in the Midwest looking at residue. That’s great, but I guarantee that if you go drive those the next year, it will look significantly different.

“Deep learning is great at interpolating conditions between what it knows; it is not good at extrapolating to situations it hasn’t seen. And in agriculture, you always feel that there is a set of conditions that you haven’t yet classified.”

A Flood of Big Data

The scale of the data is also daunting, Rostapshov points out. “We are one of the largest users of cloud computing services in the world,” he says. “We are gathering 5 to 15 million measurements per second from 130,000 connected machines globally. We have over 150 million acres in our databases, using petabytes and petabytes [of storage]. We process more data than Twitter does.”

Much of this information is so-called dirty data, that is, it doesn’t share the same format or structure, because it’s coming not only from a wide variety of John Deere machines, but also includes data from some 100 other companies that have access to the platform, including weather information, aerial imagery, and soil analyses.

As a result, says Sanchez, Deere has had to make “tremendous investments in back-end data cleanup.”

Deep learning is great at interpolating conditions between what it knows; it is not good at extrapolating to situations it hasn’t seen.”
—Julian Sanchez, John Deere

“We have gotten progressively more skilled at that problem,” he says. “We started simply by cleaning up our own data. You’d think it would be nice and neat, since it’s coming from our own machines, but there is a wide variety of different models and different years. Then we started geospatially tagging the agronomic data—the information about where you are applying herbicides and fertilizer and the like—coming in from our vehicles. When we started bringing in other data, from drones, say, we were already good at cleaning it up.”

John Deere’s Hiring Pitch

Hard problems can be a good thing to have for a company looking to hire machine learning engineers.

“Our opening line to potential recruits,” Sanchez says, “is ‘This stuff matters.’ Then, if we get a chance to talk to them more, we follow up with ‘Not only does this stuff matter, but the problems are really hard and interesting.’ When we explain the variability in farming and how we have to apply all the latest tools to these problems, we get their attention.”

Software engineers “know that feeding a growing population is a massive problem and are excited about the prospect of making a difference,” Rostapshov says.

Only 20 engineers work in the San Francisco labs right now, and that’s on a busy day—some of the researchers spend part of their time at Blue River Technology, a startup based in Sunnyvale that was acquired by Deere in 2017. About half of the researchers are focusing on AI. The Lab is in the process of doubling its office space (no word on staffing plans for that expansion yet).

“We are one of the largest users of cloud computing services in the world.”
—Alexey Rostapshov, John Deere Labs

Company-wide, Deere has thousands of software engineers, with many using AI and machine learning tools in their work, and about the same number of mechanical and electrical engineers, Sanchez reports. “If you look at our hiring 10 years ago,” he says, “it was heavily weighted to mechanical engineers. But if you look at those numbers now, it is by a large majority [engineers working] in the software space. We still need mechanical engineers—we do build green machines—but if you go by our footprint of tech talent, it is pretty safe to call John Deere a software company. And if you follow the key conversations that are happening in the company right now, 95 percent of them are software-related.”

For now, these software engineers are focused on developing technologies that allow farmers to “do more with less,” Sanchez says. Meaning, to get more and better crops from less fuel, less seed, less fertilizer, less pesticide, and fewer workers, and putting together building blocks that, he says, could eventually lead to fully autonomous farm vehicles. The data Deere collects today, for the most part, stays in silos (the virtual kind), with AI algorithms that analyze specific sets of data to provide guidance to individual farmers. At some point, however, with tools to anonymize data and buy-in from farmers, aggregating data could provide some powerful insights.

“We are not asking farmers for that yet,” Sanchez says. “We are not doing aggregation to look for patterns. We are focused on offering technology that allows an individual farmer to use less, on positioning ourselves to be in a neutral spot. We are not about selling you more seed or more fertilizer. So we are building up a good trust level. In the long term, we can have conversations about doing more with deep learning.” Continue reading

Posted in Human Robots

#435816 This Light-based Nervous System Helps ...

Last night, way past midnight, I stumbled onto my porch blindly grasping for my keys after a hellish day of international travel. Lights were low, I was half-asleep, yet my hand grabbed the keychain, found the lock, and opened the door.

If you’re rolling your eyes—yeah, it’s not exactly an epic feat for a human. Thanks to the intricate wiring between our brain and millions of sensors dotted on—and inside—our skin, we know exactly where our hand is in space and what it’s touching without needing visual confirmation. But this combined sense of the internal and the external is completely lost to robots, which generally rely on computer vision or surface mechanosensors to track their movements and their interaction with the outside world. It’s not always a winning strategy.

What if, instead, we could give robots an artificial nervous system?

This month, a team led by Dr. Rob Shepard at Cornell University did just that, with a seriously clever twist. Rather than mimicking the electric signals in our nervous system, his team turned to light. By embedding optical fibers inside a 3D printed stretchable material, the team engineered an “optical lace” that can detect changes in pressure less than a fraction of a pound, and pinpoint the location to a spot half the width of a tiny needle.

The invention isn’t just an artificial skin. Instead, the delicate fibers can be distributed both inside a robot and on its surface, giving it both a sense of tactile touch and—most importantly—an idea of its own body position in space. Optical lace isn’t a superficial coating of mechanical sensors; it’s an entire platform that may finally endow robots with nerve-like networks throughout the body.

Eventually, engineers hope to use this fleshy, washable material to coat the sharp, cold metal interior of current robots, transforming C-3PO more into the human-like hosts of Westworld. Robots with a “bodily” sense could act as better caretakers for the elderly, said Shepard, because they can assist fragile people without inadvertently bruising or otherwise harming them. The results were published in Science Robotics.

An Unconventional Marriage
The optical lace is especially creative because it marries two contrasting ideas: one biological-inspired, the other wholly alien.

The overarching idea for optical lace is based on the animal kingdom. Through sight, hearing, smell, taste, touch, and other senses, we’re able to interpret the outside world—something scientists call exteroception. Thanks to our nervous system, we perform these computations subconsciously, allowing us to constantly “perceive” what’s going on around us.

Our other perception is purely internal. Proprioception (sorry, it’s not called “inception” though it should be) is how we know where our body parts are in space without having to look at them, which lets us perform complex tasks when blind. Although less intuitive than exteroception, proprioception also relies on stretching and other deformations within the muscles and tendons and receptors under the skin, which generate electrical currents that shoot up into the brain for further interpretation.

In other words, in theory it’s possible to recreate both perceptions with a single information-carrying system.

Here’s where the alien factor comes in. Rather than using electrical properties, the team turned to light as their data carrier. They had good reason. “Compared with electricity, light carries information faster and with higher data densities,” the team explained. Light can also transmit in multiple directions simultaneously, and is less susceptible to electromagnetic interference. Although optical nervous systems don’t exist in the biological world, the team decided to improve on Mother Nature and give it a shot.

Optical Lace
The construction starts with engineering a “sheath” for the optical nerve fibers. The team first used an elastic polyurethane—a synthetic material used in foam cushioning, for example—to make a lattice structure filled with large pores, somewhat like a lattice pie crust. Thanks to rapid, high-resolution 3D printing, the scaffold can have different stiffness from top to bottom. To increase sensitivity to the outside world, the team made the top of the lattice soft and pliable, to better transfer force to mechanical sensors. In contrast, the “deeper” regions held their structure better, and kept their structure under pressure.

Now the fun part. The team next threaded stretchable “light guides” into the scaffold. These fibers transmit photons, and are illuminated with a blue LED light. One, the input light guide, ran horizontally across the soft top part of the scaffold. Others ran perpendicular to the input in a “U” shape, going from more surface regions to deeper ones. These are the output guides. The architecture loosely resembles the wiring in our skin and flesh.

Normally, the output guides are separated from the input by a small air gap. When pressed down, the input light fiber distorts slightly, and if the pressure is high enough, it contacts one of the output guides. This causes light from the input fiber to “leak” to the output one, so that it lights up—the stronger the pressure, the brighter the output.

“When the structure deforms, you have contact between the input line and the output lines, and the light jumps into these output loops in the structure, so you can tell where the contact is happening,” said study author Patricia Xu. “The intensity of this determines the intensity of the deformation itself.”

Double Perception
As a proof-of-concept for proprioception, the team made a cylindrical lace with one input and 12 output channels. They varied the stiffness of the scaffold along the cylinder, and by pressing down at different points, were able to calculate how much each part stretched and deformed—a prominent precursor to knowing where different regions of the structure are moving in space. It’s a very rudimentary sort of proprioception, but one that will become more sophisticated with increasing numbers of strategically-placed mechanosensors.

The test for exteroception was a whole lot stranger. Here, the team engineered another optical lace with 15 output channels and turned it into a squishy piano. When pressed down, an Arduino microcontroller translated light output signals into sound based on the position of each touch. The stronger the pressure, the louder the volume. While not a musical masterpiece, the demo proved their point: the optical lace faithfully reported the strength and location of each touch.

A More Efficient Robot
Although remarkably novel, the optical lace isn’t yet ready for prime time. One problem is scalability: because of light loss, the material is limited to a certain size. However, rather than coating an entire robot, it may help to add optical lace to body parts where perception is critical—for example, fingertips and hands.

The team sees plenty of potential to keep developing the artificial flesh. Depending on particular needs, both the light guides and scaffold can be modified for sensitivity, spatial resolution, and accuracy. Multiple optical fibers that measure for different aspects—pressure, pain, temperature—can potentially be embedded in the same region, giving robots a multitude of senses.

In this way, we hope to reduce the number of electronics and combine signals from multiple sensors without losing information, the authors said. By taking inspiration from biological networks, it may even be possible to use various inputs through an optical lace to control how the robot behaves, closing the loop from sensation to action.

Image Credit: Cornell Organic Robotics Lab. A flexible, porous lattice structure is threaded with stretchable optical fibers containing more than a dozen mechanosensors and attached to an LED light. When the lattice structure is pressed, the sensors pinpoint changes in the photon flow. Continue reading

Posted in Human Robots

#435791 To Fly Solo, Racing Drones Have a Need ...

Drone racing’s ultimate vision of quadcopters weaving nimbly through obstacle courses has attracted far less excitement and investment than self-driving cars aimed at reshaping ground transportation. But the U.S. military and defense industry are betting on autonomous drone racing as the next frontier for developing AI so that it can handle high-speed navigation within tight spaces without human intervention.

The autonomous drone challenge requires split-second decision-making with six degrees of freedom instead of a car’s mere two degrees of road freedom. One research team developing the AI necessary for controlling autonomous racing drones is the Robotics and Perception Group at the University of Zurich in Switzerland. In late May, the Swiss researchers were among nine teams revealed to be competing in the two-year AlphaPilot open innovation challenge sponsored by U.S. aerospace company Lockheed Martin. The winning team will walk away with up to $2.25 million for beating other autonomous racing drones and a professional human drone pilot in head-to-head competitions.

“I think it is important to first point out that having an autonomous drone to finish a racing track at high speeds or even beating a human pilot does not imply that we can have autonomous drones [capable of] navigating in real-world, complex, unstructured, unknown environments such as disaster zones, collapsed buildings, caves, tunnels or narrow pipes, forests, military scenarios, and so on,” says Davide Scaramuzza, a professor of robotics and perception at the University of Zurich and ETH Zurich. “However, the robust and computationally efficient state estimation algorithms, control, and planning algorithms developed for autonomous drone racing would represent a starting point.”

The nine teams that made the cut—from a pool of 424 AlphaPilot applicants—will compete in four 2019 racing events organized under the Drone Racing League’s Artificial Intelligence Robotic Racing Circuit, says Keith Lynn, program manager for AlphaPilot at Lockheed Martin. To ensure an apples-to-apples comparison of each team’s AI secret sauce, each AlphaPilot team will upload its AI code into identical, specially-built drones that have the NVIDIA Xavier GPU at the core of the onboard computing hardware.

“Lockheed Martin is offering mentorship to the nine AlphaPilot teams to support their AI tech development and innovations,” says Lynn. The company “will be hosting a week-long Developers Summit at MIT in July, dedicated to workshopping and improving AlphaPilot teams’ code,” he added. He notes that each team will retain the intellectual property rights to its AI code.

The AlphaPilot challenge takes inspiration from older autonomous drone racing events hosted by academic researchers, Scaramuzza says. He credits Hyungpil Moon, a professor of robotics and mechanical engineering at Sungkyunkwan University in South Korea, for having organized the annual autonomous drone racing competition at the International Conference on Intelligent Robots and Systems since 2016.

It’s no easy task to create and train AI that can perform high-speed flight through complex environments by relying on visual navigation. One big challenge comes from how drones can accelerate sharply, take sharp turns, fly sideways, do zig-zag patterns and even perform back flips. That means camera images can suddenly appear tilted or even upside down during drone flight. Motion blur may occur when a drone flies very close to structures at high speeds and camera pixels collect light from multiple directions. Both cameras and visual software can also struggle to compensate for sudden changes between light and dark parts of an environment.

To lend AI a helping hand, Scaramuzza’s group recently published a drone racing dataset that includes realistic training data taken from a drone flown by a professional pilot in both indoor and outdoor spaces. The data, which includes complicated aerial maneuvers such as back flips, flight sequences that cover hundreds of meters, and flight speeds of up to 83 kilometers per hour, was presented at the 2019 IEEE International Conference on Robotics and Automation.

The drone racing dataset also includes data captured by the group’s special bioinspired event cameras that can detect changes in motion on a per-pixel basis within microseconds. By comparison, ordinary cameras need milliseconds (each millisecond being 1,000 microseconds) to compare motion changes in each image frame. The event cameras have already proven capable of helping drones nimbly dodge soccer balls thrown at them by the Swiss lab’s researchers.

The Swiss group’s work on the racing drone dataset received funding in part from the U.S. Defense Advanced Research Projects Agency (DARPA), which acts as the U.S. military’s special R&D arm for more futuristic projects. Specifically, the funding came from DARPA’s Fast Lightweight Autonomy program that envisions small autonomous drones capable of flying at high speeds through cluttered environments without GPS guidance or communication with human pilots.

Such speedy drones could serve as military scouts checking out dangerous buildings or alleys. They could also someday help search-and-rescue teams find people trapped in semi-collapsed buildings or lost in the woods. Being able to fly at high speed without crashing into things also makes a drone more efficient at all sorts of tasks by making the most of limited battery life, Scaramuzza says. After all, most drone battery life gets used up by the need to hover in flight and doesn’t get drained much by flying faster.

Even if AI manages to conquer the drone racing obstacle courses, that would be the end of the beginning of the technology’s development. What would still be required? Scaramuzza specifically singled out the need to handle low-visibility conditions involving smoke, dust, fog, rain, snow, fire, hail, as some of the biggest challenges for vision-based algorithms and AI in complex real-life environments.

“I think we should develop and release datasets containing smoke, dust, fog, rain, fire, etc. if we want to allow using autonomous robots to complement human rescuers in saving people lives after an earthquake or natural disaster in the future,” Scaramuzza says. Continue reading

Posted in Human Robots

#435752 T-RHex Is a Hexapod Robot With ...

In Aaron Johnson’s “Robot Design & Experimentation” class at CMU, teams of students have a semester to design and build an experimental robotic system based on a theme. For spring 2019, that theme was “Bioinspired Robotics,” which is definitely one of our favorite kinds of robotics—animals can do all kinds of crazy things, and it’s always a lot of fun watching robots try to match them. They almost never succeed, of course, but even basic imitation can lead to robots with some unique capabilities.

One of the projects from this year’s course, from Team ScienceParrot, is a new version of RHex called T-RHex (pronounced T-Rex, like the dinosaur). T-RHex comes with a tail, but more importantly, it has tiny tapered toes, which help it grip onto rough surfaces like bricks, wood, and concrete. It’s able to climb its way up very steep slopes, and hang from them, relying on its toes to keep itself from falling off.

T-RHex’s toes are called microspines, and we’ve seen them in all kinds of robots. The most famous of these is probably JPL’s LEMUR IIB (which wins on sheer microspine volume), although the concept goes back at least 15 years to Stanford’s SpinyBot. Robots that use microspines to climb tend to be fairly methodical at it, since the microspines have to be engaged and disengaged with care, limiting their non-climbing mobility.

T-RHex manages to perform many of the same sorts of climbing and hanging maneuvers without losing RHex’s ability for quick, efficient wheel-leg (wheg) locomotion.

If you look closely at T-RHex walking in the video, you’ll notice that in its normal forward gait, it’s sort of walking on its ankles, rather than its toes. This means that the microspines aren’t engaged most of the time, so that the robot can use its regular wheg motion to get around. To engage the microspines, the robot moves its whegs backwards, meaning that its tail is arguably coming out of its head. But since all of T-RHex’s capability is mechanical in nature and it has no active sensors, it doesn’t really need a head, so that’s fine.

The highest climbable slope that T-RHex could manage was 55 degrees, meaning that it can’t, yet, conquer vertical walls. The researchers were most surprised by the robot’s ability to cling to surfaces, where it was perfectly happy to hang out on a slope of 135 degrees, which is a 45 degree overhang (!). I have no idea how it would ever reach that kind of position on its own, but it’s nice to know that if it ever does, its spines will keep doing their job.

Photo: CMU

T-RHex uses laser-cut acrylic legs, with the microspines embedded into 3D-printed toes. The tail is needed to prevent the robot from tipping backward.

For more details about the project, we spoke with Team ScienceParrot member (and CMU PhD student) Catherine Pavlov via email.

IEEE Spectrum: We’re used to seeing RHex with compliant, springy legs—how do the new legs affect T-RHex’s mobility?

Catherine Pavlov: There’s some compliance in the legs, though not as much as RHex—this is driven by the use of acrylic, which was chosen for budget/manufacturing reasons. Matching the compliance of RHex with acrylic would have made the tines too weak (since often only a few hold the load of the robot during climbing). It definitely means you can’t use energy storage in the legs the way RHex does, for example when pronking. T-RHex is probably more limited by motor speed in terms of mobility though. We were using some borrowed Dynamixels that didn’t allow for good positioning at high speeds.

How did you design the climbing gait? Why not use the middle legs, and why is the tail necessary?

The gait was a lot of hand-tuning and trial-and-error. We wanted a left/right symmetric gait to enable load sharing among more spines and prevent out-of-plane twisting of the legs. When using all three pairs, you have to have very accurate angular positioning or one leg pair gets pushed off the wall. Since two legs should be able to hold the full robot gait, using the middle legs was hurting more than it was helping, with the middle legs sometimes pushing the rear ones off of the wall.

The tail is needed to prevent the robot from tipping backward and “sitting” on the wall. During static testing we saw the robot tip backward, disengaging the front legs, at around 35 degrees incline. The tail allows us to load the front legs, even when they’re at a shallow angle to the surface. The climbing gait we designed uses the tail to allow the rear legs to fully recirculate without the robot tipping backward.

Photo: CMU

Team ScienceParrot with T-RHex.

What prevents T-RHex from climbing even steeper surfaces?

There are a few limiting factors. One is that the tines of the legs break pretty easily. I think we also need a lighter platform to get fully vertical—we’re going to look at MiniRHex for future work. We’re also not convinced our gait is the best it can be, we can probably get marginal improvements with more tuning, which might be enough.

Can the microspines assist with more dynamic maneuvers?

Dynamic climbing maneuvers? I think that would only be possible on surfaces with very good surface adhesion and very good surface strength, but it’s certainly theoretically possible. The current instance of T-RHex would definitely break if you tried to wall jump though.

What are you working on next?

Our main target is exploring the space of materials for leg fabrication, such as fiberglass, PLA, urethanes, and maybe metallic glass. We think there’s a lot of room for improvement in the leg material and geometry. We’d also like to see MiniRHex equipped with microspines, which will require legs about half the scale of what we built for T-RHex. Longer-term improvements would be the addition of sensors e.g. for wall detection, and a reliable floor-to-wall transition and dynamic gait transitions.

[ T-RHex ] Continue reading

Posted in Human Robots

#435703 FarmWise Raises $14.5 Million to Teach ...

We humans spend most of our time getting hungry or eating, which must be really inconvenient for the people who have to produce food for everyone. For a sustainable and tasty future, we’ll need to make the most of what we’ve got by growing more food with less effort, and that’s where the robots can help us out a little bit.

FarmWise, a California-based startup, is looking to enhance farming efficiency by automating everything from seeding to harvesting, starting with the worst task of all: weeding. And they’ve just raised US $14.5 million to do it.

FarmWise’s autonomous, AI-enabled robots are designed to solve farmers’ most pressing challenges by performing a variety of farming functions – starting with weeding, and providing personalized care to every plant they touch. Using machine learning models, computer vision and high-precision mechanical tools, FarmWise’s sophisticated robots cleanly pick weeds from fields, leaving crops with the best opportunity to thrive while eliminating harmful chemical inputs. To date, FarmWise’s robots have efficiently removed weeds from more than 10 million plants.

FarmWise is not the first company to work on large mobile farming robots. A few years ago, we wrote about DeepField Robotics and their giant weed-punching robot. But considering how many humans there are, and how often we tend to get hungry, it certainly seems like there’s plenty of opportunity to go around.

Photo: FarmWise

FarmWise is collecting massive amounts of data about every single plant in an entire field, which is something that hasn’t been possible before. Above, one of the robots at a farm in Salinas Valley, Calif.

Weeding is just one thing that farm robots are able to do. FarmWise is collecting massive amounts of data about every single plant in an entire field, practically on the per-leaf level, which is something that hasn’t been possible before. Data like this could be used for all sorts of things, but generally, the long-term hope is that robots could tend to every single plant individually—weeding them, fertilizing them, telling them what good plants they are, and then mercilessly yanking them out of the ground at absolute peak ripeness. It’s not realistic to do this with human labor, but it’s the sort of data-intensive and monotonous task that robots could be ideal for.

The question with robots like this is not necessarily whether they can do the job that they were created for, because generally, they can—farms are structured enough environments that they lend themselves to autonomous robots, and the tasks are relatively well defined. The issue right now, I think, is whether robots are really time- and cost-effective for farmers. Capable robots are an expensive investment, and even if there is a shortage of human labor, will robots perform well enough to convince farmers to adopt the technology? That’s a solid maybe, and here’s hoping that FarmWise can figure out how to make it work.

[ FarmWise ] Continue reading

Posted in Human Robots