Tag Archives: lesson

#437564 How We Won the DARPA SubT Challenge: ...

Posted on November 18, 2020 by Android

This is a guest post. The views expressed here are those of the authors and do not necessarily represent positions of IEEE or its organizational units.​

“Do you smell smoke?” It was three days before the qualification deadline for the Virtual Tunnel Circuit of the DARPA Subterranean Challenge Virtual Track, and our team was barrelling through last-minute updates to our robot controllers in a small conference room at the Michigan Tech Research Institute (MTRI) offices in Ann Arbor, Mich. That’s when we noticed the smell. We’d assumed that one of the benefits of entering a virtual disaster competition was that we wouldn’t be exposed to any actual disasters, but equipment in the basement of the building MTRI shares had started to smoke. We evacuated. The fire department showed up. And as soon as we could, the team went back into the building, hunkered down, and tried to make up for the unexpected loss of several critical hours.

Team BARCS joins the SubT Virtual Track
The smoke incident happened more than a year after we first learned of the DARPA Subterranean Challenge. DARPA announced SubT early in 2018, and at that time, we were interested in building internal collaborations on multi-agent autonomy problems, and SubT seemed like the perfect opportunity. Though a few of us had backgrounds in robotics, the majority of our team was new to the field. We knew that submitting a proposal as a largely non-traditional robotics team from an organization not known for research in robotics was a risk. However, the Virtual Track gave us the opportunity to focus on autonomy and multi-agent teaming strategies, areas requiring skill in asynchronous computing and sensor data processing that are strengths of our Institute. The prevalence of open source code, small inexpensive platforms, and customizable sensors has provided the opportunity for experts in fields other than robotics to apply novel approaches to robotics problems. This is precisely what makes the Virtual Track of SubT appealing to us, and since starting SubT, autonomy has developed into a significant research thrust for our Institute. Plus, robots are fun!

After many hours of research, discussion, and collaboration, we submitted our proposal early in 2018. And several months later, we found out that we had won a contract and became a funded team (Team BARCS) in the SubT Virtual Track. Now we needed to actually make our strategy work for the first SubT Tunnel Circuit competition, taking place in August of 2019.

Building a team of virtual robots
A natural approach to robotics competitions like SubT is to start with the question of “what can X-type robot do” and then build a team and strategy around individual capabilities. A particular challenge for the SubT Virtual Track is that we can’t design our own systems; instead, we have to choose from a predefined set of simulated robots and sensors that DARPA provides, based on the real robots used by Systems Track teams. Our approach is to look at what a team of robots can do together, determining experimentally what the best team configuration is for each environment. By the final competition, ideally we will be demonstrating the value of combining platforms across multiple Systems Track teams into a single Virtual Track team. Each of the robot configurations in the competition has an associated cost, and team size is constrained by a total cost. This provides another impetus for limiting dependence on complex sensor packages, though our ranging preference is 3D lidar, which is the most expensive sensor!

Image: Michigan Tech Research Institute

The teams can rely on realistic physics and sensors but they start off with no maps of any kind, so the focus is on developing autonomous exploratory behavior, navigation methods, and object recognition for their simulated robots.

One of the frequent questions we receive about the Virtual Track is if it’s like a video game. While it may look similar on the surface, everything under the hood in a video game is designed to service the game narrative and play experience, not require novel research in AI and autonomy. The purpose of simulations, on the other hand, is to include full physics and sensor models (including noise and errors) to provide a testbed for prototyping and developing solutions to those real-world challenges. We are starting with realistic physics and sensors but no maps of any kind, so the focus is on developing autonomous exploratory behavior, navigation methods, and object recognition for our simulated robots.

Though the simulation is more like real life than a video game, it is not real life. Due to occasional software bugs, there are still non-physical events, like the robots falling through an invisible hole in the world or driving through a rock instead of over it or flipping head over heels when driving over a tiny lip between world tiles. These glitches, while sometimes frustrating, still allow the SubT Virtual platform to be realistic enough to support rapid prototyping of controller modules that will transition straightforwardly onto hardware, closing the loop between simulation and real-world robots.

Full autonomy for DARPA-hard scenarios
The Virtual Track requirement that the robotic agents be fully autonomous, rather than have a human supervisor, is a significant distinction between the Systems and Virtual Tracks of SubT. Our solutions must be hardened against software faults caused by things like missing and bad data since our robots can’t turn to us for help. In order for a team of robots to complete this objective reliably with no human-in-the-loop, all of the internal systems, from perception to navigation to control to actuation to communications, must be able to autonomously identify and manage faults and failures anywhere in the control chain.

The communications limitations in subterranean environments (both real and virtual) mean that we need to keep the amount of information shared between robots low, while making the usability of that information for joint decision-making high. This goal has guided much of our design for autonomous navigation and joint search strategy for our team. For example, instead of sharing the full SLAM map of the environment, our agents only share a simplified graphical representation of the space, along with data about frontiers it has not yet explored, and are able to merge its information with the graphs generated by other agents. The merged graph can then be used for planning and navigation without having full knowledge of the detailed 3D map.

The Virtual Track requires that the robotic agents be fully autonomous. With no human-in-the-loop, all of the internal systems, from perception to navigation to control to actuation to communications, must be able to identify and manage faults and failures anywhere in the control chain.

Since the objective of the SubT program is to advance the state-of-the-art in rapid autonomous exploration and mapping of subterranean environments by robots, our first software design choices focused on the mapping task. The SubT virtual environments are sufficiently rich as to provide interesting problems in building so-called costmaps that accurately separate obstructions that are traversable (like ramps) from legitimately impassible obstructions. An extra complication we discovered in the first course, which took place in mining tunnels, was that the angle of the lowest beam of the lidar was parallel to the down ramps in the tunnel environment, so they could not “see” the ground (or sometimes even obstructions on the ramp) until they got close enough to the lip of the ramp to receive lidar reflections off the bottom of the ramp. In this case, we had to not only change the costmap to convince the robot that there was safe ground to reach over the lip of the ramp, but also had to change the path planner to get the robot to proceed with caution onto the top of the ramp in case there were previously unseen obstructions on the ramp.

In addition to navigation in the costmaps, the robot must be able to generate its own goals to navigate to. This is what produces exploratory behavior when there is no map to start with. SLAM is used to generate a detailed map of the environment explored by a single robot—the space it has probed with its sensors. From the sensor data, we are able to extract information about the interior space of the environment while looking for holes in the data, to determine things like whether the current tunnel continues or ends, or how many tunnels meet at an intersection. Once we have some understanding of the interior space, we can place navigation goals in that space. These goals naturally update as the robot traverses the tunnel, allowing the entire space to be explored.

Sending our robots into the virtual unknown
The solutions for the Virtual Track competitions are tested by DARPA in multiple sequestered runs across many environments for each Circuit in the month prior to the Systems Track competition. We must wait until the joint award ceremony at the conclusion of the Systems Track to find out the results, and we are completely in the dark about placings before the awards are announced. It’s nerve-wracking! The challenges of the worlds used in the Circuit events are also hand-designed, so features of the worlds we use for development could be combined in ways we have not anticipated—it’s always interesting to see what features were prioritized after the event. We test everything in our controllers well enough to feel confident that we at least are submitting something reasonably stable and broadly capable, and once the solution is in, we can’t really do anything other than “let go” and get back to work on the next phase of development. Maybe it’s somewhat like sending your kid to college: “we did our best to prepare you for this world, little bots. Go do good.”

Image: Michigan Tech Research Institute

The first SubT competition was the Tunnel Circuit, featuring a labyrinthine environment that simulated human-engineered tunnels, including hazards such as vertical shafts and rubble.

The first competition was the Tunnel Circuit, in October 2019. This environment models human-engineered tunnels. Two substantial challenges in this environment were vertical shafts and rubble. Our team accrued 21 points over 15 competition runs in five separate tunnel environments for a second place finish, behind Team Coordinated Robotics.

The next phase of the SubT virtual competition was the Urban Circuit. Much of the difference between our Tunnel and Urban Circuit results came down to thorough testing to identify failure modes and implementations of checks and data filtering for fault tolerance. For example, in the SLAM nodes run by a single robot, the coordinates of the most recent sensor data are changed multiple times during processing and integration into the current global 3D map of the “visited” environment stored by that robot. If there is lag in IMU or clock data, the observation may be temporarily registered at a default location that is very far from the actual position. Since most of our decision processes for exploration are downstream from SLAM, this can cause faulty or impossible goals to be generated, and the robots then spend inordinate amounts of time trying to drive through walls. We updated our method to add a check to see if the new map position has jumped a far distance from the prior map position, and if so, we threw that data out.

Image: Michigan Tech Research Institute

In open spaces like the rooms in the Urban circuit, we adjusted our approach to exploration through graph generation to allow the robots to accurately identify viable routes while helping to prevent forays off platform edges.

Our approach to exploration through graph generation based on identification of interior spaces allowed us to thoroughly explore the centers of rooms, although we did have to make some changes from the Tunnel circuit to achieve that. In the Tunnel circuit, we used a simplified graph of the environment based on landmarks like intersections. The advantage of this approach is that it is straightforward for two robots to compare how the graphs of the space they explored individually overlap. In open spaces like the rooms in the Urban circuit, we chose to instead use a more complex, less directly comparable graph structure based on the individual robot’s trajectory. This allowed the robots to accurately identify viable routes between features like subway station platforms and subway tracks, as well as to build up the navigation space for room interiors, while helping to prevent forays off the platform edges. Frontier information is also integrated into the graph, providing a uniform data structure for both goal selection and route planning.

The results are in!
The award ceremony for the Urban Circuit was held concurrently with the Systems Track competition awards this past February in Washington State. We sent a team representative to participate in the Technical Interchange Meeting and present the approach for our team, and the rest of us followed along from our office space on the DARPAtv live stream. While we were confident in our solution, we had also been tracking the online leaderboard and knew our competitors were going to be submitting strong solutions. Since the competition environments are hand-designed, there are always novel challenges that could be presented in these environments as well. We knew we would put up a good fight, but it was very exciting to see BARCS appear in first place!

Any time we implement a new module in our control system, there is a lot of parameter tuning that has to happen to produce reliably good autonomous behavior. In the Urban Circuit, we did not sufficiently test some parameter values in our exploration modules. The effect of this was that the robots only chose to go down small hallways after they explored everything else in their environment, which meant very often they ran out of time and missed a lot of small rooms. This may be the biggest source of lost points for us in the Urban Circuit. One of our major plans going forward from the Urban Circuit is to integrate more sophisticated node selection methods, which can help our robots more intelligently prioritize which frontier nodes to visit. By going through all three Circuit challenges, we will learn how to appropriately add weights to the frontiers based on features of the individual environments. For the Final Challenge, when all three Circuit environments will be combined into large systems, we plan to implement adaptive controllers that will identify their environments and use the appropriate optimized parameters for that environment. In this way, we expect our agents to be able to (for example) prioritize hallways and other small spaces in Urban environments, and perhaps prioritize large openings over small in the Cave environments, if the small openings end up being treacherous overall.

Next for our team: Cave Circuit
Coming up next for Team BARCS is the Virtual Cave Circuit. We are in the middle of testing our hypothesis that our controller will transition from UGVs to UAVs and developing strategies for refining our solution to handle Cave Circuit environmental hazards. The UAVs have a shorter battery life than the UGVs, so executing a joint exploration strategy will also be a high priority for this event, as will completing our work on graph sharing and merging, which will give our robot teams more sophisticated options for navigation and teamwork. We’re reaching a threshold in development where we can start increasing the “smarts” of the robots, which we anticipate will be critical for the final competition, where all of the challenges of SubT will be combined to push the limits of innovation. The Cave Circuit will also have new environmental challenges to tackle: dynamic features such as rock falls have been added, which will block previously accessible passages in the cave environment. We think our controllers are well-poised to handle this new challenge, and we’re eager to find out if that’s the case.

As of now, the biggest worries for us are time and team composition. The Cave Circuit deadline has been postponed to October 15 due to COVID-19 delays, with the award ceremony in mid-November, but there have also been several very compelling additions to the testbed that we would like to experiment with before submission, including droppable networking ‘breadcrumbs’ and new simulated platforms. There are design trade-offs when balancing general versus specialist approaches to the controllers for these robots—since we are adding UAVs to our team for the first time, there are new decisions that will have to be made. For example, the UAVs can ascend into vertical spaces, but only have a battery life of 20 minutes. The UGVs by contrast have 90 minute battery life. One of our strategies is to do an early return to base with one or more agents to buy down risk on making any artifact reports at all for the run, hedging against our other robots not making it back in time, a lesson learned from the Tunnel Circuit. Should a UAV take on this role, or is it better to have them explore deeper into the environment and instead report their artifacts to a UGV or network node, which comes with its own risks? Testing and experimentation to determine the best options takes time, which is always a worry when preparing for a competition! We also anticipate new competitors and stiffer competition all around.

Image: Michigan Tech Research Institute

Team BARCS has now a year to prepare for the final DARPA SubT Challenge event, expected to take place in late 2021.

Going forward from the Cave Circuit, we will have a year to prepare for the final DARPA SubT Challenge event, expected to take place in late 2021. What we are most excited about is increasing the level of intelligence of the agents in their teamwork and joint exploration of the environment. Since we will have (hopefully) built up robust approaches to handling each of the specific types of environments in the Tunnel, Urban, and Cave circuits, we will be aiming to push the limits on collaboration and efficiency among the agents in our team. We view this as a central research contribution of the Virtual Track to the Subterranean Challenge because intelligent, adaptive, multi-robot collaboration is an upcoming stage of development for integration of robots into our lives.

The Subterranean Challenge Virtual Track gives us a bridge for transitioning our more abstract research ideas and algorithms relevant to this degree of autonomy and collaboration onto physical systems, and exploring the tangible outcomes of implementing our work in the real world. And the next time there’s an incident in the basement of our building, the robots (and humans) of Team BARCS will be ready to respond.

Richard Chase, Ph.D., P.E., is a research scientist at Michigan Tech Research Institute (MTRI) and has 20 years of experience developing robotics and cyber physical systems in areas from remote sensing to autonomous vehicles. At MTRI, he works on a variety of topics such as swarm autonomy, human-swarm teaming, and autonomous vehicles. His research interests are the intersection of design, robotics, and embedded systems.

Sarah Kitchen is a Ph.D. mathematician working as a research scientist and an AI/Robotics focus area leader at MTRI. Her research interests include intelligent autonomous agents and multi-agent collaborative teams, as well as applications of autonomous robots to sensing systems.

This material is based upon work supported by the Defense Advanced Research Projects Agency (DARPA) under Contract No. HR001118C0124 and is released under Distribution Statement (Approved for Public Release, Distribution Unlimited). Any opinions, findings and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of DARPA. Continue reading →

Posted in Human Robots

#436944 Is Digital Learning Still Second Best?

Posted on March 20, 2020 by Android

As Covid-19 continues to spread, the world has gone digital on an unprecedented scale. Tens of thousands of employees are working from home, and huge conferences, like the Google I/O and Apple WWDC software extravaganzas, plan to experiment with digital events.

Universities too are sending students home. This might have meant an extended break from school not too long ago. But no more. As lecture halls go empty, an experiment into digital learning at scale is ramping up. In the US alone, over 100 universities, from Harvard to Duke, are offering online classes to students to keep the semester going.

While digital learning has been improving for some time, Covid-19 may not only tip us further into a more digitally connected reality, but also help us better appreciate its benefits. This is important because historically, digital learning has been viewed as inferior to traditional learning. But that may be changing.

The Inversion
We often think about digital technologies as ways to reach people without access to traditional services—online learning for children who don’t have schools nearby or telemedicine for patients with no access to doctors. And while these solutions have helped millions of people, they’re often viewed as “second best” and “better than nothing.” Even in more resource-rich environments, there’s an assumption one should pay more to attend an event in person—a concert, a football game, an exercise class—while digital equivalents are extremely cheap or free. Why is this? And is the situation about to change?

Take the case of Dr. Sanjeev Arora, a professor of medicine at the University of New Mexico. Arora started Project Echo because he was frustrated by how many late-stage cases of hepatitis C he encountered in rural New Mexico. He realized that if he had reached patients sooner, he could have prevented needless deaths. The solution? Digital learning for local health workers.

Project Echo connects rural healthcare practitioners to specialists at top health centers by video. The approach is collaborative: Specialists share best practices and work through cases with participants to apply them in the real world and learn from edge cases. Added to expert presentations, there are lots of opportunities to ask questions and interact with specialists.

The method forms a digital loop of learning, practice, assessment, and adjustment.

Since 2003, Project Echo has scaled to 800 locations in 39 countries and trained over 90,000 healthcare providers. Most notably, a study in The New England Journal of Medicine found that the outcomes of hepatitis C treatment given by Project Echo trained healthcare workers in rural and underserved areas were similar to outcomes at university medical centers. That is, digital learning in this context was equivalent to high quality in-person learning.

If that is possible today, with simple tools, will they surpass traditional medical centers and schools in the future? Can digital learning more generally follow suit and have the same success? Perhaps. Going digital brings its own special toolset to the table too.

The Benefits of Digital
If you’re training people online, you can record the session to better understand their engagement levels—or even add artificial intelligence to analyze it in real time. Ahura AI, for example, founded by Bryan Talebi, aims to upskill workers through online training. Early study of their method suggests they can significantly speed up learning by analyzing users’ real-time emotions—like frustration or distraction—and adjusting the lesson plan or difficulty on the fly.

Other benefits of digital learning include the near-instantaneous download of course materials—rather than printing and shipping books—and being able to more easily report grades and other results, a requirement for many schools and social services organizations. And of course, as other digitized industries show, digital learning can grow and scale further at much lower costs.

To that last point, 360ed, a digital learning startup founded in 2016 by Hla Hla Win, now serves millions of children in Myanmar with augmented reality lesson plans. And Global Startup Ecosystem, founded by Christine Souffrant Ntim and Einstein Kofi Ntim in 2015, is the world’s first and largest digital accelerator program. Their entirely online programs support over 1,000 companies in 90 countries. It’s astonishing how fast both of these organizations have grown.

Notably, both examples include offline experiences too. Many of the 360ed lesson plans come with paper flashcards children use with their smartphones because the online-offline interaction improves learning. The Global Startup Ecosystem also hosts about 10 additional in-person tech summits around the world on various topics through a related initiative.

Looking further ahead, probably the most important benefit of online learning will be its potential to integrate with other digital systems in the workplace.

Imagine a medical center that has perfect information about every patient and treatment in real time and that this information is (anonymously and privately) centralized, analyzed, and shared with medical centers, research labs, pharmaceutical companies, clinical trials, policy makers, and medical students around the world. Just as self-driving cars can learn to drive better by having access to the experiences of other self-driving cars, so too can any group working to solve complex, time-sensitive challenges learn from and build on each other’s experiences.

Why This Matters
While in the long term the world will likely end up combining the best aspects of traditional and digital learning, it’s important in the near term to be more aware of the assumptions we make about digital technologies. Some of the most pioneering work in education, healthcare, and other industries may not be highly visible right now because it is in a virtual setting. Most people are unaware, for example, that the busiest emergency room in rural America is already virtual.

Once they start converging with other digital technologies, these innovations will likely become the mainstream system for all of us. Which raises more questions: What is the best business model for these virtual services? If they start delivering better healthcare and educational outcomes than traditional institutions, should they charge more? Hopefully, we will see an even bigger shift occurring, in which technology allows us to provide high quality education, healthcare, and other services to everyone at more affordable prices than today.

These are some of the topics we can consider as Covid-19 forces us into uncharted territory.

Image Credit: Andras Vas / Unsplash Continue reading →

Posted in Human Robots

#436526 Not Bot, Not Beast: Scientists Create ...

Posted on January 22, 2020 by Android

A remarkable combination of artificial intelligence (AI) and biology has produced the world’s first “living robots.”

This week, a research team of roboticists and scientists published their recipe for making a new lifeform called xenobots from stem cells. The term “xeno” comes from the frog cells (Xenopus laevis) used to make them.

One of the researchers described the creation as “neither a traditional robot nor a known species of animal,” but a “new class of artifact: a living, programmable organism.”

Xenobots are less than 1 millimeter long and made of 500-1,000 living cells. They have various simple shapes, including some with squat “legs.” They can propel themselves in linear or circular directions, join together to act collectively, and move small objects. Using their own cellular energy, they can live up to 10 days.

While these “reconfigurable biomachines” could vastly improve human, animal, and environmental health, they raise legal and ethical concerns.

Strange New ‘Creature’
To make xenobots, the research team used a supercomputer to test thousands of random designs of simple living things that could perform certain tasks.

The computer was programmed with an AI “evolutionary algorithm” to predict which organisms would likely display useful tasks, such as moving towards a target.

After the selection of the most promising designs, the scientists attempted to replicate the virtual models with frog skin or heart cells, which were manually joined using microsurgery tools. The heart cells in these bespoke assemblies contract and relax, giving the organisms motion.

The creation of xenobots is groundbreaking. Despite being described as “programmable living robots,” they are actually completely organic and made of living tissue. The term “robot” has been used because xenobots can be configured into different forms and shapes, and “programmed” to target certain objects, which they then unwittingly seek. They can also repair themselves after being damaged.

Possible Applications
Xenobots may have great value. Some speculate they could be used to clean our polluted oceans by collecting microplastics. Similarly, they may be used to enter confined or dangerous areas to scavenge toxins or radioactive materials. Xenobots designed with carefully shaped “pouches” might be able to carry drugs into human bodies.

Future versions may be built from a patient’s own cells to repair tissue or target cancers. Being biodegradable, xenobots would have an edge on technologies made of plastic or metal.

Further development of biological “robots” could accelerate our understanding of living and robotic systems. Life is incredibly complex, so manipulating living things could reveal some of life’s mysteries—and improve our use of AI.

Legal and Ethical Questions
Conversely, xenobots raise legal and ethical concerns. In the same way they could help target cancers, they could also be used to hijack life functions for malevolent purposes.

Some argue artificially making living things is unnatural, hubristic, or involves “playing God.” A more compelling concern is that of unintended or malicious use, as we have seen with technologies in fields including nuclear physics, chemistry, biology and AI. For instance, xenobots might be used for hostile biological purposes prohibited under international law.

More advanced future xenobots, especially ones that live longer and reproduce, could potentially “malfunction” and go rogue, and out-compete other species.

For complex tasks, xenobots may need sensory and nervous systems, possibly resulting in their sentience. A sentient programmed organism would raise additional ethical questions. Last year, the revival of a disembodied pig brain elicited concerns about different species’ suffering.

Managing Risks
The xenobot’s creators have rightly acknowledged the need for discussion around the ethics of their creation. The 2018 scandal over using CRISPR (which allows the introduction of genes into an organism) may provide an instructive lesson here. While the experiment’s goal was to reduce the susceptibility of twin baby girls to HIV-AIDS, associated risks caused ethical dismay. The scientist in question is in prison.

When CRISPR became widely available, some experts called for a moratorium on heritable genome editing. Others argued the benefits outweighed the risks.

While each new technology should be considered impartially and based on its merits, giving life to xenobots raises certain significant questions:

Should xenobots have biological kill-switches in case they go rogue?
Who should decide who can access and control them?
What if “homemade” xenobots become possible? Should there be a moratorium until regulatory frameworks are established? How much regulation is required?

Lessons learned in the past from advances in other areas of science could help manage future risks, while reaping the possible benefits.

Long Road Here, Long Road Ahead
The creation of xenobots had various biological and robotic precedents. Genetic engineering has created genetically modified mice that become fluorescent in UV light.

Designer microbes can produce drugs and food ingredients that may eventually replace animal agriculture. In 2012, scientists created an artificial jellyfish called a “medusoid” from rat cells.

Robotics is also flourishing. Nanobots can monitor people’s blood sugar levels and may eventually be able to clear clogged arteries. Robots can incorporate living matter, which we witnessed when engineers and biologists created a sting-ray robot powered by light-activated cells.

In the coming years, we are sure to see more creations like xenobots that evoke both wonder and due concern. And when we do, it is important we remain both open-minded and critical.

This article is republished from The Conversation under a Creative Commons license. Read the original article.

Image Credit: Photo by Joel Filipe on Unsplash Continue reading →

Posted in Human Robots

#435522 Harvard’s Smart Exo-Shorts Talk to the ...

Posted on August 20, 2019 by Android

Exosuits don’t generally scream “fashionable” or “svelte.” Take the mind-controlled robotic exoskeleton that allowed a paraplegic man to kick off the World Cup back in 2014. Is it cool? Hell yeah. Is it practical? Not so much.

Yapping about wearability might seem childish when the technology already helps people with impaired mobility move around dexterously. But the lesson of the ill-fated Google Glassholes, which includes an awkward dorky head tilt and an assuming voice command, clearly shows that wearable computer assistants can’t just work technologically—they have to look natural and allow the user to behave like as usual. They have to, in a sense, disappear.

To Dr. Jose Pons at the Legs + Walking Ability Lab in Chicago, exosuits need three main selling points to make it in the real world. One, they have to physically interact with their wearer and seamlessly deliver assistance when needed. Two, they should cognitively interact with the host to guide and control the robot at all times. Finally, they need to feel like a second skin—move with the user without adding too much extra mass or reducing mobility.

This week, a US-Korean collaboration delivered the whole shebang in a Lululemon-style skin-hugging package combined with a retro waist pack. The portable exosuit, weighing only 11 pounds, looks like a pair of spandex shorts but can support the wearer’s hip movement when needed. Unlike their predecessors, the shorts are embedded with sensors that let them know when the wearer is walking versus running by analyzing gait.

Switching between the two movement modes may not seem like much, but what naturally comes to our brains doesn’t translate directly to smart exosuits. “Walking and running have fundamentally different biomechanics, which makes developing devices that assist both gaits challenging,” the team said. Their algorithm, computed in the cloud, allows the wearer to easily switch between both, with the shorts providing appropriate hip support that makes the movement experience seamless.

To Pons, who was not involved in the research but wrote a perspective piece, the study is an exciting step towards future exosuits that will eventually disappear under the skin—that is, implanted neural interfaces to control robotic assistance or activate the user’s own muscles.

“It is realistic to think that we will witness, in the next several years…robust human-robot interfaces to command wearable robotics based on…the neural code of movement in humans,” he said.

A “Smart” Exosuit Hack
There are a few ways you can hack a human body to move with an exosuit. One is using implanted electrodes inside the brain or muscles to decipher movement intent. With heavy practice, a neural implant can help paralyzed people walk again or dexterously move external robotic arms. But because the technique requires surgery, it’s not an immediate sell for people who experience low mobility because of aging or low muscle tone.

The other approach is to look to biophysics. Rather than decoding neural signals that control movement, here the idea is to measure gait and other physical positions in space to decipher intent. As you can probably guess, accurately deciphering user intent isn’t easy, especially when the wearable tries to accommodate multiple gaits. But the gains are many: there’s no surgery involved, and the wearable is low in energy consumption.

Double Trouble
The authors decided to tackle an everyday situation. You’re walking to catch the train to work, realize you’re late, and immediately start sprinting.

That seemingly easy conversion hides a complex switch in biomechanics. When you walk, your legs act like an inverted pendulum that swing towards a dedicated center in a predictable way. When you run, however, the legs move more like a spring-loaded system, and the joints involved in the motion differ from a casual stroll. Engineering an assistive wearable for each is relatively simple; making one for both is exceedingly hard.

Led by Dr. Conor Walsh at Harvard University, the team started with an intuitive idea: assisted walking and running requires specialized “actuation” profiles tailored to both. When the user is moving in a way that doesn’t require assistance, the wearable needs to be out of the way so that it doesn’t restrict mobility. A quick analysis found that assisting hip extension has the largest impact, because it’s important to both gaits and doesn’t add mass to the lower legs.

Building on that insight, the team made a waist belt connected to two thigh wraps, similar to a climbing harness. Two electrical motors embedded inside the device connect the waist belt to other components through a pulley system to help the hip joints move. The whole contraption weighed about 11 lbs and didn’t obstruct natural movement.

Next, the team programmed two separate supporting profiles for walking and running. The goal was to reduce the “metabolic cost” for both movements, so that the wearer expends as little energy as needed. To switch between the two programs, they used a cloud-based classification algorithm to measure changes in energy fluctuation to figure out what mode—running or walking—the user is in.

Smart Booster
Initial trials on treadmills were highly positive. Six male volunteers with similar age and build donned the exosuit and either ran or walked on the treadmill at varying inclines. The algorithm performed perfectly at distinguishing between the two gaits in all conditions, even at steep angles.

An outdoor test with eight volunteers also proved the algorithm nearly perfect. Even on uneven terrain, only two steps out of all test trials were misclassified. In an additional trial on mud or snow, the algorithm performed just as well.

“The system allows the wearer to use their preferred gait for each speed,” the team said.

Software excellence translated to performance. A test found that the exosuit reduced the energy for walking by over nine percent and running by four percent. It may not sound like much, but the range of improvement is meaningful in athletic performance. Putting things into perspective, the team said, the metabolic rate reduction during walking is similar to taking 16 pounds off at the waist.

The Wearable Exosuit Revolution
The study’s lightweight exoshorts are hardly the only players in town. Back in 2017, SRI International’s spin-off, Superflex, engineered an Aura suit to support mobility in the elderly. The Aura used a different mechanism: rather than a pulley system, it incorporated a type of smart material that contracts in a manner similar to human muscles when zapped with electricity.

Embedded with a myriad of sensors for motion, accelerometers and gyroscopes, Aura’s smartness came from mini-computers that measure how fast the wearer is moving and track the user’s posture. The data were integrated and processed locally inside hexagon-shaped computing pods near the thighs and upper back. The pods also acted as the control center for sending electrical zaps to give the wearer a boost when needed.

Around the same time, a collaboration between Harvard’s Wyss Institute and ReWalk Robotics introduced a fabric-based wearable robot to assist a wearer’s legs for balance and movement. Meanwhile, a Swiss team coated normal fabric with electroactive material to weave soft, pliable artificial “muscles” that move with the skin.

Although health support is the current goal, the military is obviously interested in similar technologies to enhance soldiers’ physicality. Superflex’s Aura, for example, was originally inspired by technology born from DARPA’s Warrior Web Program, which aimed to reduce a soldier’s mechanical load.

That said, military gear has had a long history of trickling down to consumer use. Similar to the way camouflage, cargo pants, and GORE-TEX trickled down into the consumer ecosphere, it’s not hard to imagine your local Target eventually stocking intelligent exowear.

Image and Video Credit: Wyss Institute at Harvard University. Continue reading →

Posted in Human Robots

#434585 This Week’s Awesome Stories From ...

Posted on February 2, 2019 by Android

ARTIFICIAL INTELLIGENCE
The World’s Fastest Supercomputer Breaks an AI Record
Tom Simonite | Wired
“Summit, which occupies an area equivalent to two tennis courts, used more than 27,000 powerful graphics processors in the project. It tapped their power to train deep-learning algorithms, the technology driving AI’s frontier, chewing through the exercise at a rate of a billion billion operations per second, a pace known in supercomputing circles as an exaflop.”

ROBOTICS
iRobot Finally Announces Awesome New Terra Robotic Lawnmower
Evan Ackerman | IEEE Spectrum
“Since the first Roomba came out in 2002, it has seemed inevitable that one day iRobot would develop a robotic lawn mower. After all, a robot mower is basically just a Roomba that works outside, right? Of course, it’s not nearly that simple, as iRobot has spent the last decade or so discovering, but they’ve finally managed to pull it off.”

3D Printing
Watch This Super Speedy 3D Printer Make Objects Suddenly Appear
Erin Winick | MIT Technology Review
“The new machine—which the team nicknamed the ‘replicator’ after the machine from Star Trek—instead forms the entire item all in one go. It does this by shining light onto specific spots in a rotating resin that solidifies when exposed to a certain light level.”

GENETICS
The DIY Designer Baby Project Funded With Bitcoin
Antonio Regalado | MIT Technology Review
“i‘Is DIY bio anywhere close to making a CRISPR baby? No, not remotely,’ David Ishee says. ‘But if some rich guy pays a scientist to do the work, it’s going to happen.’ He adds: ‘What you are reporting on isn’t Bryan—it’s the unseen middle space, a layer of gray-market biotech and freelance science where people with resources can get things done.’i”

SCIENCE
The Complete Cancer Cure Story Is Both Bogus and Tragic
Megan Molteni | Wired
“You’d think creators and consumers of news would have learned their lesson by now. But the latest version of the fake cancer cure story is even more flagrantly flawed than usual. The public’s cancer cure–shaped amnesia, and media outlets’ willingness to exploit it for clicks, are as bottomless as ever. Hope, it would seem, trumps history.”

BOOKS
An AI Reading List—From Practical Primers to Sci-Fi Short Stories
James Vincent | The Verge
“The Verge has assembled a reading list: a brief but diverse compendium of books, short stories, and blogs, all chosen by leading figures in the AI world to help you better understand artificial intelligence.”

Image Credit: Katya Havok / Shutterstock.com Continue reading →

Posted in Human Robots