Innovative design: Mab deploys flying robots to clean house

Twenty-three-year-old Colombian designer Adrian Perez Zapata designed a home-cleaning system that has won the top prize at the Electrolux Design Lab competition. According to Zapata, his design was inspired by the controlled flight of bees.

The technology, called The Mab, is a circular automated cleaning system that scans an entire house for areas that need to be cleaned, then deploys hundreds of tiny, flying robots to carry out the task. The Mab comes with a total of 908 robots, which are solar powered and each clean using one drop of water.

The main docking device itself is equipped with a filter to keep reusable water clean, and pills which are deposited into the system through a capsule give water viscosity and surface tension to help the robots clean more efficiently.

- Robojentics (Robotics), Sweeya Virtu Hub, Rajahmundry.

Tiny Robot Pulls Objects 2,000 Times Its Size

A tiny robot that weighs no more than a bouncy ball can pull objects nearly 2,000 times its size — the equivalent of a human pulling a blue whale, according to researchers at Stanford University.

This robot, called a MicroTug, can also hoist an object 100 times its weight vertically up a glass surface, which researchers say is the equivalent of a human climbing up a skyscraper while carrying an elephant.

DNews: Rescue Robots Battle It Out At DARPA Challenge

At Stanford University’s Biomimetics and Dextrous Manipulation Lab, a team of researchers looked to nature when designing the miniature robots and emulated the uniquely grippy feet of ants and geckos that help them climb surfaces.

As a result, the MicroTug uses strong directional adhesives to tightly grip a surface. The robot, equipped with a large winch, tow cable and hook, then uses its onboard motors to pull the objects. The researchers use 3-D printers to make the parts.

In the video below, the 12g MicroTug pulls off some pretty impressive feats, like pulling a 600g coffee mug. The miniature robots suggest that adhesion can dramatically outperform friction robots.

Robot Ants Mimic A Real Colony

The project’s researchers see the miniature robots, which only cost about $20 to make, as a launching pad for bigger things to come. They are now looking at ways to make multiple MicroTugs work together as a team and then plan to scale the technology up to larger robots that have much more force.

In the future, researchers say we could see these types of robots pulling objects out of hard-to-reach spots or even sent into a burning building to find survivors.

Underwater robot swarms use collective cognition to perform tasks

Scientists have created underwater robot swarms that function like schools of fish, exchanging information to monitor the environment, searching, maintaining, exploring and harvesting resources in underwater habitats. The EU supported COCORO project explored and developed collective cognition in autonomous robots in a rich set of 10 experimental demonstrators, which are shown in the 52 videos.

The COCORO project's robot swarms not only look like schools of fish, they behave like them too. The project developed autonomous robots that interact with each other and exchange information, resulting in a cognitive system that is aware of its environment.

According to Dr. Thomas Schmickl, coordinator of the project and Associate Professor in the Department of Zoology at the University of Graz in Austria, what distinguishes COCORO from other similar projects is that researchers created robot swarms that are capable of collective cognition. They work as a collective system of autonomous agents that can learn from past experience and their environment.

Robot swarm cognition in action

In one experiment, twenty Jeff robots floated in a tank of water. As they came into contact with each other, they gradually became aware of the size of their swarm. This 'swarm size awareness' was made possible by relaying status information using LEDs.

In another scenario, the robots' mission was to find debris originating from a sunken airplane. Lily robots searched just below the surface while Jeff robots searched at the bottom of the pool.

Magnets were placed around the airplane to mimic an electro-magnetic signal emitted locally and the robots used their built-in compasses to locate the target. A Jeff robot soon discovered the target and settled on it at the bottom of the pool.

By transmitting LED, it 'recruited' the other Jeff robots, which then gathered around the target, while Lily robots collected overhead.

During field trials in Livorno Harbour, Italy, the robots were exposed to waves, currents and corrosive salt water. Despite the difficult conditions the robot swarms were able to remain clustered around their base station as well as go on "patrols" and successfully return to base.

Bio-mimicry: inspired by nature

'We didn't invent all of this ourselves,' says Dr. Schmickl, explaining that COCORO's scientists modelled collective cognition in nature. Observing how honeybees cluster, for example, helped them to develop the BEECLUST algorithm that they used to aggregate robots at a specific location. They also applied mechanisms derived from existing studies on how slime mould amoebas congregate using chemical waves to communicate with each other.

A diverse group of biologists, computer scientists and other experts participated in COCORO, which ran from 1 April 2011 until 30 September 2014 and received EUR 2.9 million in EU funding.

Although the project concluded in 2014, its results could have wide application in the fields of computer science, biology, theology, meta-cognition, psychology, and philosophy, as well as a broader impact on our economy and society. Possible applications are in distributed environmental monitoring and search&rescue operations.

- Robojentics ( Robotics ), Sweeya Virtu Hub, Rajahmundry.

The World’s First Personal Robot

Welcome to a new time with your personal robots Luna. It is the world’s first personal robot, which is announced to have the ability to handle almost all kinds of works and even comes with its own app store where additional functionalities such as face recognition would be sold.

It is also built with 8-inch touchscreen, two cameras, wireless connectivity, a three-mic array, and a variety of sensors. Coming soon before the end of 2011, and you may reserve one now.

- Robojentics (Robotics), Sweeya Virtu Hub, Rajahmundry.

Google’s self-driving cars are now cruising the streets of California!

Like hoverboards and jetpacks, self-driving cars are one of those 'too-good-to-be-true' technologies that could change everything for everyone living in a big city. Except unlike hoverboards and jetpacks - you’ve got a lot to prove, Lexus! - self-driving cars are an actual reality right now, and are starting to make real changes to how the residents of Mountain View in California get around.

As Mike Murphy points out at Quartz, these things look like adorable little pandas, and they’re getting Californians where they need to go, no driver’s license required.Quartz reports that they’re only travelling at 40 km/hour right now, and have a driver behind the wheel at all times in case of an emergency. As we reported back in May, a review found that one in 12 self-driving cars have been in accidents since September 2014, but humans were at fault for the collision, not the cars.

Unfortunately though, Google’s self-driving cars still have their training wheels firmly attached -

By getting them out on the public roads now, Google is aiming to increase the 1.6 million km of experience that their self-driving cars so far have under their very round belts. More driving means more data on what could possibly go wrong for these cars when they’re interacting with regular cars. While the crashes over the past 16 months weren't serious and caused no injuries (and no wonder - they were travelling at less than 16 km/hour at the time) researchers at Google are keen to figure out exactly what kinds of situations will put them in danger, and use that information to further improve their cars' AI.

And together with their own data, they’d like to gather information from people who have actually encountered one of their self-driving cars on the streets too. Chris Ziegler reports for The Verge that Google has launched a website where drivers can recount their experiences of having to interact with one of these sweet-looking bubble-mobiles, telling them if it went smoothly, if it required extra concentration to move around the car, or if it was just kinda weird sharing the streets with something not controlled by a human brain. 

"It'll probably help that some of the cars are going to become rolling works of art - how can you hate a cute car covered in an even cuter pattern?” says Ziegler. "Then again, the cars are capped at 'a neighbourhood-friendly 25 mph (40 km/h),' so if you've ever wanted to lay on your horn at an artificially intelligent vehicle that's going way too slow on a public street, this might be a wonderful opportunity."

Just everyone please be patient with these plucky little robots. They're trying their best to create a world where less and less of us are compelled to spend our valuable hours behind the wheel, let's hope they get us there. 

- Robojentics, Sweeya Virtu Hub, Rajahmundry.

The 20 jobs that robots are most likely to take over By Kathleen Elkins

Machines are only getting smarter and more efficient.

So much so that they’re starting to take over both blue-collar and white-collar jobs.

NPR recently posted a guide created by researchers that predicts the chance of our jobs being automated within the next 20 years.

To make their projections, the experts scored jobs across 21 fields on nine possible traits, the four most important being: cleverness, negotiation, helping others, and squeezing into small spaces.

Mental health and substance abuse social workers appear to be in the clear, with a 0.3% chance of being automated.

Telemarketers, umpires, cashiers, and several other jobs that are already being replaced by robots, have reason to worry, however. You can interact with the full guide here.

While the researchers admitted that these percentages are rough, the data offers insights into what the future may look like.

20. Electrical and electronic equipment assemblers

Electrical and electronic equipment assemblers have a 95.1% chance of being automated.

Robots are becoming critical to the production process of electronics because of their precision and accuracy. They are being used to do things such as: load solar wafers into solar cells, place LED light bulbs, and inspect circuit boards, as reported by the Robotics Industries Association.

19. Postal service workers

Postal service workers have a 95.4% chance of being automated.

Postal sorters, clerks, and mail carriers are being hit hard by automation. Not only are robots able to do tasks such as sort mail, but snail mail is also becoming more and more obsolete with the increasing digitization of mail.

18. Jewelers and precious stone and metal workers

Jewelers and precious stone and metal workers have a 95.5% chance of being automated.

The Bureau of Labor Statistics predicts their employment to decline 10% between now and 2022 as robots begin assisting in the manufacturing and repairing of jewelry.

17. Restaurant cooks

Restaurant cooks have a 96.3% chance of being automated.

A noodle-slicing robot named Foxbot can be found at Dazzling Noodles, an open-kitchen restaurant chain in North China’s Shanxi province.

There’s another robot chef making crab bisque from scratch, thanks to 20 motors, 24 joints, and 129 sensors. The robot, designed by Moley Robotics, can complete the complicated dish in 30 minutes and even plates it.

16. Grinding and polishing workers

Workers who grind and polish materials have a 97% chance of being automated.

Robots are increasingly able to grind or polish a variety of metal, wood, stone, clay, plastic, and glass objects.

15. Cashiers

Cashiers have a 97.1% chance of being automated.

Panera Bread announced that it will have replaced all of its cashiers with kiosks by 2016, according to a USA Today article.

Additionally, self-checkout machines are continuing to spring up in grocery stores around the world.

14. Bookkeepers

Bookkeepers have a 97.6% chance of being automated.

The Wall Street Journal reported that big companies such as Pilot Travel, Verizon, and GameStop, among others, are using software to automate corporate bookkeeping tasks.

These companies now only need about 10 clerks to pay suppliers, as opposed to the 80 or so they would need without the robot bookkeepers.

13. Legal secretaries

Legal secretaries have a 97.6% chance of being automated.

Secretaries are disappearing from the job market now that technology allows bosses to field calls and arrange meetings themselves.

12. Fashion models

Surprisingly, fashion models have a 97.6% chance of being automated.

In addition to “taking over” fashion model jobs, robots are performing in other surprising ways, such as acting. Engineered Arts, a British company, has created a fully interactive and multilingual robot called the RoboThespian, which can hold eye contact, guess a person’s mood and age, and break into song.

11. Drivers

Drivers have a 97.8% chance of being automated.

Drivers and chauffeurs won’t be needed for much longer. Google’s self-driving test cars have driven thousands of miles without human intervention. Also, Uber CEO Travis Kalanick loves the idea of autonomous vehicles and announced that Uber will eventually be replacing all of its drivers with cars that drive themselves.

10. Credit analysts

Credit analysts have a 97.9% chance of being automated.
Automating the process of analyzing credit data and financial statements, and preparing reports with credit information, could result in a lower degree of risk.

9. Milling and planing machine setters, operators, and tenders

Milling and planing machine setters, operators, and tenders have a 97.9% chance of being automated.

Setting up, operating, or tending milling or planing machines are tasks that are increasingly being delegated to machines.

8. Packaging and filling-machine operators and tenders

Packaging and filling machine operators and tenders have a 98% chance of being automated.
Using robots to prepare industrial or consumer products for storage or shipment is becoming more and more prevalent. Amazon now uses an army of robots — each 320 pounds and 16 inches tall — to bring shelves of goods out of storage in order to be shipped, reported Time.

7. Procurement clerks

Procurement clerks have a 98% chance of being automated.
It is now very simple for machines to place orders with suppliers for materials and services. Also, ordering over the internet — or “e-procurement” — will result in a decline of these jobs.

6. Umpires and referees

Umpires and referees have a 98.3% chance of being automated.
In professional tennis, a computerized umpire called Hawk Eye is already being used to help the chief umpire make close line calls. Players have the option of “challenging” a call, in which the Hawk Eye system will display where the ball landed and whether it was in or out. Its decisions are final.

Hawk Eye represents just one system being used to solve disputes in sports.

5. Tellers

Tellers have a 98.3% chance of being automated.
Chances are you haven’t used a human bank teller in a while. ATMs can provide most of the services that tellers offer.

4. Loan officers

Loan officers have a 98.4% chance of being automated.
A Bloomberg article reported that inroads are already being made at Daric Inc., an online peer-to-peer lender. The company has replaced all loan officers with an algorithm that identifies safe borrowers.

3. Timing-device assemblers and adjusters

Timing device assemblers and adjusters have a 98.5% chance of being automated.Machines are now able to perform the precise assembling, adjusting, or calibrating that timing device assemblers specialize in.

2. Tax preparers

Tax preparers have a 98.7% chance of being automated.

Automating the process of preparing tax returns could result in much fewer errors, and the technology, Optical Character Recognition (OCR), already exists.

1. Telemarketers

Telemarketers have a 99% chance of being automated.
Many of today’s cold-callers are not human. Robots can not only perform the job 24/7, but they can also maintain energy and perkiness no matter how many rude consumers they interact with.

- Robojentics, Sweeya Virtu Hub, Rajahmundry.

Pepper first emotional humanoid robot to be sold in Japan for $1,600

A joint venture between Japan’s SoftBank, China’s Alibaba and Foxconn Technology is bringing the emotional human-like robot Pepper to the global market. It’s the first step to make robots available to general consumers for home and work.

Always Be Chatty: Japanese robot Pepper will sell you a coffee machine… with a smile
Foxconn and Alibaba are each investing $118 million (¥14.5 billion) for a stake of about 20 percent each in the company, with SoftBank holding the remaining 60 percent.

Pepper is the world's first personal robot and will go on sale in Japan this weekend, with monthly production of about 1,000 units, according to SoftBank. Sales outside Japan will not begin earlier than next year. The 1.2-meter tall robot will sell for $1,600, but there will be additional monthly fees of about $200 to use it, including an insurance package.

“Robots are going to be as popular as cars,” Alibaba’s founder Jack Ma said in Tokyo on Thursday during his appearance with SoftBank’s CEO Masayoshi Son. “They will be everywhere and I think we all have to get ready for that.”

Pepper will be the first robot to react to human emotions and even have emotions of its own. He moves on wheels and looks like is wearing an iPad as a pendant. He takes information from his cameras and touch sensors, and his behavior changes based on how this makes him feel. The humanoid robot will be provided with some 200 apps, such as voice recognition though he won’t be able to do dishes or laundry. SoftBank hopes that robot will grow emotionally with its owners and become a family member so they won’t become bored with it.

“Robots have already surpassed human beings in calculation and memory, but I have no doubt that the time will come when they will surpass in wisdom as well,” Softbank’s head, Japanese billionaire Masayoshi Son said. While Jack Ma added, that the only key challenge was adding “the heart,” which now sets humans apart from machines.

During the demonstration on Thursday, Pepper who was introduced to the world last year, danced, took photos and seemed to demonstrate different emotions.

SoftBank also plans to launch several models for enterprises, 'Pepper for Biz,' this autumn. The company wants its robots to serve as baby-sitters, medical workers or even party companions in the future. Robots are already being used at stores including SoftBank's own mobile phone shops.

Softbank, Alibaba and Foxconn are striving to spread and develop the robotics industry on a worldwide scale as they want to become key players in the sector.

- Robojentics (Robotics), Sweeya Virtu Hub, Rajahmundry.

Korean Robot Takes Home $2M Prize in DARPA Challenge.

A robotics team from South Korea took home the $2 million first-place prize in a competition this weekend to design robots that could aid humans in a natural or man-made disaster.

During the DARPA Robotics Challenge Finals, which took place here Friday and Saturday (June 5 and 6), the winning team's DRC-HUBO robot finished all eight tasks in less than 45 minutes. The winning bot had a humanoid design that could transform itself into a wheeled kneeling position for faster, more stable movement.

The Running Man robot from Florida's Institute for Human and Machine Cognition claimed second, finishing all the tasks in just over 50 minutes. Team Tartan Rescue's CHIMP robot came in third, completing all the tasks in just over 55 minutes. The second and third place teams took home $1 million and $500,000, respectively.

"Today was incredible. It was everything we hoped it would be and more," Gill Pratt, the DARPA program manager in charge of the challenge, said in a news conference today. But robotics is still in a "young age," he said.

The U.S. Defense Advanced Research Projects Agency started the challenge in 2012. The competition was inspired by the Fukushima Daiichi nuclear disaster, caused by the deadly 2011 earthquake and tsunami in Japan. The idea was to develop robots that could aid emergency responders in disaster recovery efforts by going places that are unsafe for humans.

The 25 teams had two chances to complete as many as possible of the course's eight tasks, which included driving a utility vehicle, exiting the vehicle, opening a door, cutting a hole in a wall, walking over a pile of rubble (or clearing a path through debris), walking up a short flight of stairs, and completing a surprise task, which was different each day.

The bots had to complete the course without a tether to prevent them from falling over and despite a disrupted communications link between the bots and their human operators.

After the first day of competition, Carnegie Mellon University's CHIMP robot (short for CMU Highly Intelligent Mobile Platform) was in the lead, becoming the first team to complete all eight tasks in the hour allotted. CHIMP was followed by the German Team NIMBRO RESCUE's robot Momaro and the Jet Propulsion Laboratory's RoboSimian robot, which scored seven points each during the first day.

But the second day of competition brought an upset by South Korea's team KAIST, whose robot leaped into first place after finishing all eight tasks in less time than CMU's CHIMP robot — including the surprise task, which involved unplugging a wire and plugging it into another outlet. Team IHMC also completed all eight tasks on the second day to claim second place, bumping CHIMP into third.

Though the field of robotics still has a long way to go, the teams made some impressive strides in robotic technology. But even more surprising was the way the audience reacted to the robots, Pratt said. "We have robots that look like the Terminator," he said, and yet spectators fell in love with them, cheering when the robots succeeded and feeling sorry for them when they failed. "There is some incredible, untapped affinity for machines," Pratt said.

Top 10 team rankings:

1. TEAM KAIST (8 points, 44:28 minutes)

2. TEAM IHMC ROBOTICS (8 points, 50:26 minutes)

3. TARTAN RESCUE (8 points, 55:15 minutes)

4. TEAM NIMBRO RESCUE (7 points, 34:00)

5. TEAM ROBOSIMIAN (7 points, 47:59 minutes)

6. TEAM MIT (7 points, 50:25 minutes)

7. TEAM WPI-CMU (7 points, 56:06 minutes)

8. TEAM DRC-HUBO @ UNLV (6 points, 57:41 minutes)

9. TEAM TRAC LABS (5 points, 49:00 minutes)

10. TEAM AIST-NEDO (5 points, 52:30 minutes)

- Robojentics (Robotics), Sweeya Virtu Hub, Rajahmundry

MIT’s Robotic Cheetah Can Jump Over High Obstacles By Itself!

After making their headless cheetahs able to roam autonomously around the lab, the MIT robots have now been given a considerable upgrade in the form of leaping over obstacles. A LIDAR system helps it achieve those moves. The cheetah had been jumping over obstacles up to the height of 33cm, but now it will be able to do much more than that!

Leaping robots might seem very simple in appearance, but it takes a lot of calculations and judgment to do such a move. The obstacle measuring system has to measure the exact height, length and width of the obstacle above the surface. Then it can do necessary calculations for the right pose to jump and bend while in mid-air. The reason behind the Cheetah robot project is that the world’s fastest runner and hunter hold great lessons for the future of robotics. There are various other organizations working on such cheetahs like Darpa and Boston Dynamics.

But only the cheetah from MIT can claim autonomous jumping ability over obstacles. With the help of onboard LIDAR system, the machine is now able to use Laser reflections to map the terrain. A special parallel algorithm dictates the following moves involved in the jumping process. The first part involves the measuring the size and distance of the obstacle. The second part of the programming sees the robot change direction and speed to have the best approach possible. This hard part only takes one-tenth of a second and dynamically changes the robot’s working shape. But that is not all. A third part of the algorithm involves working out the best working trajectory possible. The calculation of the precise force required to make the perfect leap is its part.

The cheetah was untethered on an indoor test track it performed admirably well, jumping over 45 cm obstacles with 90 percent success rate. It doesn’t even need to stop before attempting a jump. This dynamic achievement to handle balance and impact after leaping is a commendable effort indeed. But even now, the robot needs firm ground to walk on. It can’t successfully make a jump over soft terrain like grass. It limits its functionality outdoors, so it can’t be called a fully functional cheetah as of yet. The team is working hard to get it outdoors and so are the others. Which one will be the first to come up with the best all-terrain cheetah? We will have to wait and see!

- Robojenticcs (Robotics) - Sweeya Virtu Hub, Rajahmundry.

Google has purchased the cutting-edge robotics firm

Google has purchased the cutting-edge robotics firm which supplies mobile research robots for the Pentagon. The tech giant is keeping secret what it will produce with the acquired technology.

Boston Dynamics, an engineering company that designs robots like BigDog, Cheetah, WildCat, and Atlas, is now the eighth robotic company to join Google's ranks in the last six months, the company confirmed Friday, according to The New York Times.

The robotics company - based in Waltham, Massachusetts - is known for its fascinating robots that have a sense of balance and can walk and run on almost any kind of terrain. Boston Dynamics was established in 1992 and is known for supplying robotics technology mostly for Pentagon clients, including the Defense Advanced Research Projects Agency (DARPA).

Google has so far refused to reveal what it is planning to do with the new technology, but the internet giant's robotic efforts, headed by Andy Rubin, will certainly have impressive capability due to the new acquisition. The company also said it will keep the financial aspect of the deal confidential.

The tech giant said that it intends to honor the existing military contracts the firm has, but does not plan to become a military contractor.
As of now, Boston Dynamics has a $10.8 million contract with DARPA to supply several humanoid Atlas robots. The machines are set to participate in the DARPA Robotics Challenge, which awards the winning maker with a $2 million prize. The contest evaluates robots' ability to function during natural disasters and catastrophes such as an earthquake or a nuclear power meltdown. The company's robots are famous for their abilities to walk through terrain which is difficult even for humans. One of its four-legged robots, BigDog, can climb hills, walk through snow, travel through ice, and remain standing even after being kicked by a human, according to a video of the robot released in 2008. One of the company's recent creations is WildCat, a four-legged outdoor runner capable of rising, turning, and reaching running speeds of up to 16 miles per hour (mph) on flat ground. The robotics firm has also designed robots that can climb walls and trees, as well as run faster than any human being on earth. A video recorded the Cheetah robot running 29mph faster than the fastest human, Usain Bolt. The Jamaican sprinter's top speed is 27.78 mph (44.7km/h). Other robotics companies purchased by Google are based in the US and Japan, and focus on software for advanced robotic arms, grasping technology, and computer vision.

- Robojentics (Robotics) - Sweeya Virtu Hub, Rajahmundry.

Robots can recover from damage in minutes!

Robots will one day provide tremendous benefits to society, such as in search and rescue missions and putting out forest fires - but not until they can learn to keep working if they become damaged. A new paper in the journal Nature, called "Robots That Can Adapt Like Animals," shows how to make robots automatically recover from injury in less than two minutes.

A video of the work shows a six-legged robot that adapts to keep walking even if two of its legs are broken. It also shows a robotic arm that learned how to correctly place an object even with several broken motors.

Antoine Cully and Jean-Baptiste Mouret, from the Pierre and Marie Curie University in France, led the work in collaboration with Jeff Clune (University of Wyoming) and Danesh Tarapore (Pierre and Marie Curie University).

In contrast to today's robots, animals exhibit an amazing ability to adapt to injury. There are many three-legged dogs that can catch Frisbees, for example, and if your ankle is sprained, you quickly figure out a way to walk despite the injury. The scientists took inspiration from these biological strategies.

"When injured, animals do not start learning from scratch," senior author Jean-Baptiste Mouret said. "Instead, they have intuitions about different ways to behave. These intuitions allow them to intelligently select a few, different behaviors to try out and, after these tests, they choose one that works in spite of the injury. We made robots that can do the same."

Before it is deployed, the robot uses a computer simulation of itself to create a detailed map of the space of high-performing behaviors. This map represents the robot's "intuitions" about different behaviors it can perform and their predicted value. If the robot is damaged, it uses these intuitions to guide a learning algorithm that conducts experiments to rapidly discover a compensatory behavior that works despite the damage. The new algorithm is called "Intelligent Trial and Error."

"Once damaged, the robot becomes like a scientist," explains lead author Antoine Cully. "It has prior expectations about different behaviors that might work, and begins testing them. However, these predictions come from the simulated, undamaged robot. It has to find out which of them work, not only in reality, but given the damage.

"Each behavior it tries is like an experiment and, if one behavior doesn't work, the robot is smart enough to rule out that entire type of behavior and try a new type," Cully continues. "For example, if walking, mostly on its hind legs, does not work well, it will next try walking mostly on its front legs. What's surprising is how quickly it can learn a new way to walk. It's amazing to watch a robot go from crippled and flailing around to efficiently limping away in about two minutes."

The same Intelligent Trial and Error algorithm allows robots to adapt to unforeseen situations, including adapting to new environments and inventing new behaviors.

Jeff Clune explains that "technically, Intelligent Trial and Error involves two steps: (1) creating the behavior-performance map, and (2) adapting to an unforeseen situation."

The map in the first step is created with a new type of evolutionary algorithm called MAP-Elites. Evolutionary algorithms simulate Darwinian evolution by hosting "survival of the fittest" competitions in computer simulations to evolve artificially intelligent robots. The adaptation in the second step involves a "Bayesian optimization" algorithm that takes advantage of the prior knowledge provided by the map to efficiently search for a behavior that works despite the damage.

"We performed experiments that show that the most important component of Intelligent Trial and Error is creating and harnessing the prior knowledge contained in the map," Clune says.

This new technique will help develop more robust, effective, autonomous robots. Danesh Tarapore provides some examples.

"It could enable the creation of robots that can help rescuers without requiring their continuous attention," he says. "It also makes easier the creation of personal robotic assistants that can continue to be helpful even when a part is broken."

- Robojentics  (Robotics) - Sweeya Virtu Hub, Rajahmundry.