CMU develops a method to improve robotic grasping of transparent objects

Picking up transparent objects is hard when you’re a robot. Many of the traditional cameras and sensors just can’t get a good enough view to tell the grasper where to go. The light from infrared cameras passes through the objects and gets scattered in the process, and depth cameras have trouble determine the proper shape without opaque surfaces.

The result of all of this is a high failure rate as robot hands attempt to graph into something tactile. It can be a particular issue among those looking to employ robots for recycling, as so many plastic and glass bottles are clear.

Carnegie Mellon this week issued some new research that could refine the process, using standard consumer cameras. In this case, the team of researchers created a color camera that’s able to determine the shape of a transparent object based on color readings. It’s still imperfect — and not as accurate as with opaque objects — but the researchers say they can grasp the clear surfaces with a much higher rate of success than previous methods.

“We do sometimes miss,” CMU assistant robotics professor David Held said in a release tied to the announcement,  “but for the most part it did a pretty good job, much better than any previous system for grasping transparent or reflective objects.”

More of the findings will be reported in a virtual robotics conference later this summer.

White Castle becomes the first fast food chain to test out the robot fry cook, Flippy, from Miso Robotics

The next time Harold and Kumar go to a White Castle, there may be a robot making their French Fries.

In one of the first trials of a robotic fry cook at a national burger chain, White Castle said it would work with Pasadena, Calif.-based Miso Robotics to test that company’s robotic chef at a restaurant in the Chicago area. It’s a  trial run for potentially bringing the robot to other White Castle kitchens across the country, the company said.

White Castle first began talking about using the Miso Robotics robots in its kitchens about nine months ago according to White Castle’s vice president of shareholder relations, Jamie Richardson. For the company, it was a question of, “How can we start to make the kitchen of tomorrow today?” 

Already a success on social media, where videos of Miso Robotics’ Flippy robot have racked up hundreds of thousands of views, White Castle was intrigued about the prospects of a burger flipping, chicken, onion, and french frying robot in its locations, Richardson said.

“I think automation is here to stay and this is the first example of a really large credible player starting down that journey,” said Miso Robotics chief executive Buck Jordan of the new collaboration with White Castle. 

White Castle has a fairly interesting track record when it comes to working with startups. The company was the first fast food chain to embrace Impossible Foods for its sliders.

At an undisclosed restaurant in the Chicago area, Miso Robotics is already working to install the latest version of its Flippy robot. The robotic fry cook will be integrated with the company’s point of sale system so that the robot can begin preparation as soon as an order is taken at the register.

That first robot will be coming online in September, according to Richardson.

And Richardson said that White Castle employees don’t need to worry about a robot coming for all of their jobs… yet. 

“It’s going to save us money in food costs because there will be less waste,” said Richardson.  “The other savings will be in terms of output… that’s going to be helpful.. If you maintain speed of service that’s getting a little bit better and a little better you do see more visits… that’s where we see it having the biggest impact… we’re not looking at this as a way to reduce people power.” 

A typical installation of a Miso Robotics system in a kitchen would cost a restaurant $30,000 upfront and then another $15,000 per year. However, with White Castle, the terms (which were undisclosed) were a little different.

Jordan said the goal is to bring the cost of the robotic system down to $15,000 for the entire system, obviating the need for any upfront costs, and convincing restaurants and franchisors that the robot can pay for itself right out of the gate.

There’s a clear path to getting that down to 20K,” said Jordan. “I’m trying to chisel that down to 15K,… at that kind of price and these things have lifetimes of seven to ten years we can afford to take the loss upfront.”

The robots have taken on new significance in the post COVID-19 era as restaurants like White Castle become essential services even as they struggle to keep the lights on with fewer customers. 

At White Castle that meant pay cuts for executives in order to retain staff. “We cut a lot of investment and we didn’t want to lose one job,” Richardson said. However, even with the strategic cuts, the implementation of at least this first robotic system remained a priority.

“There were things that we thought, COVID or no COVID were important,” Richardson said. “This project falls under that banner.”

White Castle’s decision to pilot Flippy in the kitchen creates an avenue for reduced human contact with food during the cooking process – reducing potential for transmission of food pathogens. The implementation also brings intelligence to cooking, tapping into sensors, intelligent monitoring and anticipated kitchen needs to keep food temperatures consistent, that ensure optimal quality and a perfect bite for customers. With Flippy in the kitchen automating repetitive, time consuming and dangerous tasks like frying, team members can be redeployed to more customer-experience driven tasks.

Image Credit: Miso Robotics

Garry Kasparov on AI: ‘People always called me an optimist’

Garry Kasparov is a political activist who’s written books and articles on artificial intelligence, cybersecurity and online privacy, but he’s best known for being the former World Chess Champion who took on the IBM computer known as Big Blue in the mid-1990s.

I spoke to Kasparov before a speaking engagement at the Collision Conference last month where he was participating in his role as Avast Security Ambassador. Our discussion covered a lot of ground, from his role as security ambassador to the role of AI. (Transcribed questions and answers were edited for clarity.)

TechCrunch: How did you become a security ambassador for Avast?

Garry Kasparov: It started almost by accident. I was invited by one of my friends, who knew the previous Avast CEO (Vince Steckler) to be the guest speaker at the opening of their new headquarters in Prague. I met the team and very quickly we recognized that we could work together very effectively since Avast wanted an ambassador.

I thought that it would be a great combination because it’s about cybersecurity, and it’s also about customers, about individual rights, which is related to human rights, and it also had a little bit of a political element of course. But most importantly, it’s a combination of privacy and security and I felt that with my record of working for human rights, and also writing about individuals and privacy and also having some experience with computers, that it would be a good match.

Now it’s my fourth year and it seems that many of the things we have been discussing at conferences when I have spoken about the role of AI in our lives, and many of the discussions that we thought were theoretical, have become more practical.

What were those discussions like?

One of the favorite topics that was always raised at these conferences is whether AI will be a helping hand or threat. And my view has been that it’s neither because I have always said that AI was neither a magic wand nor a Terminator. It’s a tool. And it’s up to us to find the best way of using it and applying its enormous power to our good.

Festo’s latest biomimetic robots are a flying feathered bird and ball-bottomed helper arm

You could be excused for thinking that German robotics company Festo does nothing but put together fabulous prototype robots built to resemble kangaroos, jellyfish, and other living things. They do in fact actually make real industrial robots, but it’s hard not to marvel at their biomimetic experiments; Case in point, the feathered BionicSwift and absurd BionicMobileAssistant motile arm.

Festo already has a flying bird robot — I wrote about it almost 10 years ago. They even made a flying bat as a follow-up. But the BionicSwift is more impressive than both because, in an effort to more closely resemble its avian inspiration, it flies using artificial feathers.

Image Credits: Festo

“The individual lamellae [i.e. feathers] are made of an ultralight, flexible but very robust foam and lie on top of each other like shingles. Connected to a carbon quill, they are attached to the actual hand and arm wings as in the natural model,” Festo writes in its description of the robot.

The articulating lamellae allow the wing to work like a bird’s, forming a powerful scoop on the downstroke to push against the air, but separating on the upstroke to produce less resistance. Everything is controlled on-board, including the indoor positioning system that the bird was ostensibly built to demonstrate. Flocks of BionicSwifts can fly in close quarters and avoid each other using an ultra wideband setup.

Festo’s BionicMobileAssistant seems like it would be more practical, and in a way it is, but not by much. The robot is basically an arm emerging from a wheeled base — or rather a balled one. The spherical bottom is driven by three “omniwheels,” letting it move easily in any direction while minimizing its footprint.

The hand is a showcase of modern robotic gripper design, with all kinds of state of the art tech packed in there — but the result is less than the sum of its parts. What makes a robotic hand good these days is less that it has a hundred sensors in the palm and fingers and huge motility for its thumb, but rather intelligence about what it is gripping. An unadorned pincer may be a better “hand” than one that looks like the real thing because of the software that backs it up.

Not to mention the spherical movement strategy makes for something of an unstable base. It’s telling that the robot is transporting scarves and not plates of food or parts.

Of course, it’s silly to criticize such a machine, which is aspirational rather than practical. But it’s important to understand that these fascinating creations from Festo are hints at a possible future more than anything.

China’s Pudu raises $15M for indoor delivery robots

The robotics category has been building to a kind of critical mass in recent years, but the past six months of the COVID-19 pandemic have pushed many otherwise wary investors over the top. Today, Shenzhen-based Pudu Robotics announced that it has completed a $15 million Series B, with Beijing food services group, Meituan as the sole investor.

Pudu describes itself as a “smart delivery robotics” company, with a majority of its products falling within the food services category. There are multiple robotic SKUs for food delivery and dish return, all of which are indoor models. Rather than focusing on delivery apps, the robotics are designed for a variety of scenarios under the same roof, including hotels, restaurants and office buildings.

Last month, Pudu noted that it has deployed “hundreds” of units to hospitals in South Korea and its native China amid the pandemic. Other existing clients include restaurants and hotels, all of which are looking for methods for reducing human contact as a means of transmitting the novel coronavirus. In total, it says its robots have been deployed in 200 cities across 20 countries.

“Non-human physical contact means safety, and automation means saving human efforts. In the event of human life, these two advantages will be magnified,” CEO Zang Tao said in a press release issued last month. “Many technology companies have played an important role in intelligent disinfection, unmanned delivery and intelligent diagnosis during COVID-19, which made an irreversible influence to the public health system.”

The “irreversible” bit remains to be seen, of course. What does seems certain, however, is that COVID-19 will be in important testing ground for the efficacy and need for these sorts of technologies. What seemed like, at best, an indulgence a year ago is now being viewed as a potentially necessary part of the food handling process. The virus has certainly driven investor interest, but it will be up to the startups to show they’re really to deliver on the promise.

Elementary Robotics is making its quality assurance robots commercially available

Two years and over $17 million after it first began working on its robots for quality assurance, the Los Angeles-based Elementary Robotics has finally made its products commercially available.

The company already boasts a few very large initial customers in the automotive industry, consumer packaged goods, and aerospace and defense, including Toyota, according to chief executive Arye Barnehama. Now, the robotics technology that Barnehama and his co-workers have been developing for years is broadly available to other companies beyond its six initial pilot customers.

The company’s robots look like a large box with a gantry system providing three degrees of freedom, with vertical and horizontal movement as well as a gimbal-mounted camera that can visualize products.

Image credit: Elementary Robotics

As objects are scanned by the robots they’re compared against a taxonomy of objects provided by the companies that Elementary works with to determine whether or not there’s a defect.

Barnehama also emphasizes that Elementary’s robots are not designed to replace every human interaction or assessment in the manufacturing process. “Machine learning paired with humans always performs better,” says Barnehama. “At the end of the day the human is running the factory. We’re not really a lights out factory.”

Behind the new commercialization push is a fresh $12.7 million in financing that Elementary closed at the end of 2019.

The lead investor in that round was Threshold Ventures and the firm’s partner, Mo Islam, has already taken a seat on the Elementary Robotics board of directors, while existing investors Fika Ventures, Fathom Capital and Toyota AI Ventures, also participated in the round, which will be used to allow Elementary Robotics to continue developing and deploying its automation products at scale, the company said.

“Robotics and particularly robotics applied to manufacturing has been an interest of mine,” said Islam. In Elementary Robotics, Islam saw a company that could compete with large, publicly traded businesses like Cognex. The low complexity and ease of deployment of Elementary’s hardware was another big selling point for Islam that convinced him to invest. 

Elementary says that it can be up and running at a site in a matter of days and with businesses emphasizing cost-cutting and enabling remote work to ensure worker safety, companies are embracing the technology.

“That’s where we’re really excited to be launching it,” said Barnehama. “If we get parts or data examples we can get that up and running same day. We can usually show customers within that week we can start showing them the value of that as we get more and more data through the system.”

Cities are wrestling with a potential new exodus in the COVID-19 era, but Urban-X still believes in their future

Embodying the tensions that cities across the world face as they wrestle with controlling a pandemic in dense, urban environments, Urban-X, the accelerator for technology startups focused on the problems cities face, has launched its eighth, fully remote, cohort.

While the accelerator program backed by the BMW-owned Mini Cooper automaker and the venture capital firm Urban Us is based in Brooklyn, it’s conducting its latest program virtually, with participating startups coming from Atlanta, Sydney, San Francisco, Boston, Burlington, and Los Angeles, according to a statement.

“Long term, we are bullish on cities. I think that COVID and climate change share some things in common. If we think that COVID is disruptive, and not only a threat to economic livelihood but human life, climate change, is certainly a much larger threat,” said Micah Kotch, the managing director of Urban-X. “I think that cities have withstood pandemics previously. I think that we will moving forward. The clear things that we need are really good political leadership. We need to heed science and to act quickly based on the best possible science and we need collective action. And that’s where I see a lot of overlap between covid and climate.”

The latest batch of companies that Urban-X will work with includes:

  • Adiona: a machine learning-based service to optimize hourly workforces in logistics and supply chain management. 
  • Aquagenuity: a company providing search information about water quality for consumers
  • Climate Robotics: a manufacturer of robots that produce carbon-sequestering and soil-improving biochar
  • Mobilyze: the developer of a data analytics service for electric vehicle charging station site optimization
  • Resonant Link: the creator of a wireless charging service to power robots and electric vehicles.
  • Xtelligent: a company rethinking traffic signal technology

“Not everyone can afford to move out to the suburbs and not everyone wants to. Cities are going to continue to be the epicenters of creativity and innovation,” said Kotch. “While these last three-and-a-half to four months have been a real challenge, particularly here in the U.S. we are deep believers in the vibrancy and necessity of cities.”

Later stage investors think that the Urban-X thesis can create viable businesses, with about 85 percent of the accelerator’s companies going on to raise additional rounds of funding. Some of the most successful companies (in terms of capital raised) include Bowery Farming, Starcity, Mark43, One Concern, Future Motion, Skycatch, Seamlessdocs, Revivn, BRCK and Rachio.

“Technology, investment and mentorship have the power to advance the low carbon, resilient and high density future we need for our cities,” said Shaun Abrahamson, URBAN-X Investment Committee and Managing Partner at Urban Us. “We are thrilled to have this new group of founders join URBAN-X to build creative solutions that tackle climate change and the biggest issues facing our cities.”

Geek+, the Amazon Kiva of China, lands $200 million Series C

Geek+, a Beijing-based startup that makes warehouse fulfillment robots similar to those of Amazon’s Kiva, said Thursday that it has closed over $200 million in a Series C funding round.

That bumps total capital raised by the 5-year-old company to date to nearly $390 million. The new round, completed in two parts, was separately led by GGV Capital and D1 Capital Partners in the summer of 2019, and V Fund earlier this year. Other participants included Warburg Pincus, Redview Capital and Vertex Ventures.

The company said it will continue to develop robotics solutions tailored to logistics, ramp up its robot-as-a-service monetization model, and expand partnerships.

Geek+ represents a rank of Chinese robotics solution providers that are increasingly appealing to clients around the world. The startup itself boasts over 10,000 robots deployed worldwide, serving 300 customers and projects in over 20 countries.

Just last month, Geek+ announced a partnership with Conveyco, an order fulfillment and distribution center system integrator operating in North America, to distribute its autonomous mobile robots (ARMs) across the continent. Leading this part of its business is Mark Messina, the startup’s chief operating officer for the Americas who previously worked at Amazon, where he oversaw mechanical engineering for the Kiva robotics system.

Geek+’s ambitious move overseas came amid continuous pressure from the Trump administration to boycott Chinese tech players. Back home, Geek+ has worked closely with retail giants such as Alibaba and Suning to replace human pickers in warehouses.

Software will reshape our world in the next decade

As I was wrapping up a Zoom meeting with my business partners, I could hear my son joking with his classmates in his online chemistry class.

I have to say this is a very strange time for me: As much as I love my family, in normal times, we never spend this much time together. But these aren’t normal times.

In normal times, governments, businesses and schools would never agree to shut everything down. In normal times, my doctor wouldn’t agree to see me over video conferencing.

No one would stand outside a grocery store, looking down to make sure they were six feet apart from one another. In times like these, decisions that would normally take years are being made in a matter of hours. In short, the physical world — brick-and-mortar reality— has shut down. The world still functions, but now it is operating inside everyone’s own home.

This not-so-normal time reminds me of 2008, the depths of the financial crisis. I sold my company BEA Systems, which I co-founded, to Oracle for $8.6 billion in cash. This liquidity event was simultaneously the worst and most exhausting time of my career, and the best time of my career, thanks to the many inspiring entrepreneurs I was able to meet.

These were some of the brightest, hardworking, never-take-no-for-an-answer founders, and in this era, many CEOs showed their true colors. That was when Slack, Lyft, Uber, Credit Karma, Twilio, Square, Cloudera and many others got started. All of these companies now have multibillion dollar market caps. And I got to invest and partner with some of them.

Once again, I can’t help but wonder what our world will look like in 10 years. The way we live. The way we learn. The way we consume. The way we will interact with each other.

What will happen 10 years from now?

Welcome to 2030. It’s been more than two decades since the invention of the iPhone, the launch of cloud computing and one decade since the launch of widespread 5G networks. All of the technologies required to change the way we live, work, eat and play are finally here and can be distributed at an unprecedented speed.

The global population is 8.5 billion and everyone owns a smartphone with all of their daily apps running on it. That’s up from around 500 million two decades ago.

Robust internet access and communication platforms have created a new world.

The world’s largest school is a software company — its learning engine uses artificial intelligence to provide personalized learning materials anytime, anywhere, with no physical space necessary. Similar to how Apple upended the music industry with iTunes, all students can now download any information for a super-low price. Tuition fees have dropped significantly: There are no more student debts. Kids can finally focus on learning, not just getting an education. Access to a good education has been equalized.

The world’s largest bank is a software company and all financial transactions are digital. If you want to talk to a banker live, you’ll initiate a text or video conference. On top of that, embedded fintech software now powers all industries.

No more dirty physical money. All money flow is stored, traceable and secured on a blockchain ledger. The financial infrastructure platforms are able to handle customers across all geographies and jurisdictions, all exchanges of value, all types of use-cases (producers, distributors, consumers) and all from the start.

The world’s largest grocery store is a software and robotics company — groceries are delivered whenever and wherever we want as fast as possible. Food is delivered via robot or drones with no human involvement. Customers can track where, when and who is involved in growing and handling my food. Artificial intelligence tells us what we need based on past purchases and our calendars.

The world largest hospital is a software and robotics company — all initial diagnoses are performed via video conferencing. Combined with patient medical records all digitally stored, a doctor in San Francisco and her artificial intelligence assistant can provide personalized prescriptions to her patients in Hong Kong. All surgical procedures are performed by robots, with supervision by a doctor of course, we haven’t gone completely crazy. And even the doctors get to work from home.

Our entire workforce works from home: Don’t forget the main purpose of an office is to support companies’ workers in performing their jobs efficiently. Since 2020, all companies, and especially their CEOs, realized it was more efficient to let their workers work from home. Not only can they save hours of commute time, all companies get to save money on office space and shift resources toward employee benefits. I’m looking back 10 years and saying to myself, “I still remember those days when office space was a thing.”

The world’s largest entertainment company is a software company, and all the content we love is digital. All blockbuster movies are released direct-to-video. We can ask Alexa to deliver popcorn to the house and even watch the film with friends who are far away. If you see something you like in the movie, you can buy it immediately — clothing, objects, whatever you see — and have it delivered right to your house. No more standing in line. No transport time. Reduced pollution. Better planet!

These are just a few industries that have been completely transformed by 2030, but these changes will apply universally to almost anything. We were told software was eating the world.

The saying goes you are what you eat. In 2030, software is the world.

Security and protection no longer just applies to things we can touch and see. What’s valuable for each and every one of us is all stored digitally — our email account, chat history, browsing data and social media accounts. It goes on and on. We don’t need a house alarm, we need a digital alarm.

Even though this crisis makes the near future seem bleak, I am optimistic about the new world and the new companies of tomorrow. I am even more excited about our ability to change as a human race and how this crisis and technology are speeding up the way we live.

This storm shall pass. However the choices we make now will change our lives forever.

My team and I are proud to build and invest in companies that will help shape the new world; new and impactful technologies that are important for many generations to come, companies that matter to humanity, something that we can all tell our grandchildren about.

I am hopeful.

No-code industrial robotics programming startup Wandelbots raises $30 million

Dresden, Germany-based Wandelbots – a startup dedicated to making it easier for non-programmers to ‘teach’ industrial robots how to do specific tasks – has raised a $30 million Series B funding round led by 83North, an with participation from Next47 and Microsoft’s M12 venture funding arm.

Wandelbots will use the funding to help it speed the market debut of its TracePen, a hand-held, code-free device that allows human operators to quickly and easily demonstrate desired behavior for industrial robots to mimic. Programming robots to perform specific tasks typically requires massive amounts of code, as well as programmers with very specific, in-demand skillsets to accomplish; Wandelbots wants to make it as easy as simply showing a robot what it is you want it to do – and then showing it a different set of behaviors should you need to reprogram it to accomplish a new task or fill in for a different part of the assembly line.

The software that Wandelbots developed to make this possible originally sprung out of work done at the Faculty of Computer Science at the Technical University of Dresden. The startup was a finalist in our TechCrunch Disrupt Battlefield competition in 2017, and raised a $6.8 million Series A round in 2018 led by Paua Ventures, EQT Ventures and others.

Wandelbots already has some big-name clients, including industrial giants like Volkswagen, BMW, Infineon and others, and as of June 17, it’ll be launching its TracePen publicly for the first time. The company’s technology has the potential to save anyone who makes use of industrial robots many months of programming time, and the associated costs – and could ultimately make use of this kind of robotics practical even for smaller companies for whom the budgetary requirements of doing so previously put it out of reach.

I asked Wandelbots CEO and co-founder Christian Piechnick via email whether their platform can overcome some of the challenges companies including Tesla have faced with introducing ever-greater automation to their manufacturing facilities.

“The reversals regarding automation were caused by the inflexibility, complexity and cost introduced by automation with robots,” Piechnick told me via email. “People are usually not aware that 75% of the total cost of ownership of a robot comes from software development. The problems introduced by robots were killing the benefit. This is exactly the problem we are tackling. We enable manufacturers to use robots with an unseen flexibility and we dramatically lower the cost of using robots. Our product enables non-programmers to easily teach a robot new tasks and thus, reduces the involvement of hard-to-find and costly programmers.”

TracePen, the device and companion platform that Wandelbots is launching this week, is actually an evolution of their original vision, which focuses more on using smart clothes to fully model human behavior in real-time in order to translate that to robotic instruction. The company’s pivot to TracePen employs the same underlying software tech, but meets customers much closer to where they already are in terms of processes and operations, while still providing the same cost reduction benefits and flexibility, according to Piechnick.

I asked Piechnick about COVID-19 and how that has impacted Wandelbots’ business, and he replied that in fact it’s driven up demand for automation, and efficiencies that benefit automation, in a number of key ways.

“COVID-19 has impacted the thinking on global manufacturing in various ways,” he wrote. “First there is the massive trend of reshoring to reduce the risk of globally distributed supply chains. In order to scale volume, ensure quality and reduce cost, automation is a natural consequence for developed countries. With a technology that leads to almost immediate ROI and extremely short time-to-market, we hit a trend. Furthermore, the dependency on human workers and the workplace restrictions (e.g., distance between workers) increases the demand for automation tremendously.”