Artificial Intelligence and Robotics blog
Awesome-o
This user hasn't shared any biographical information
Posts by Awesome-o
MABEL two-legged robot fastest in the world
Aug 22nd
It would appear that we have a new champion in the “what robot can run fastest race”. The two-legged robot MABEL under development for several years at the University of Michigan was recently revealed to reach a top running speed of 6.8 miles per hour or roughly 11 kilometers per hour. This means that MABEL is significantly faster than the previous record holder which was Toyota’s humanoid robot with a top speed of 7 kilometers per hour; Honda’s ASIMO is now in 3rd place with a top speed of 6 kilometers per hour.
Another technical achievement behind MABEL and above Toyota’s robot is that MABEL’s gate while walking and running much more closely resembles that of a human. In addition, MABEL jumps 3-4 inches above ground (both legs) when running whereas Toyota’s robot hardly does (it does enough to consider it’s fast walking gate running). One disadvantage is that MABEL is not a complete robot that includes an upper body with arms and a head compared to Toyota’s, Honda’s and other humanoid robots coming out of Korea and Japan.
The below video shows MABEL running including some explanation of the related technical achievement. Additional information on the project can be found at the project’s website here.
Robotic seagull takes flight in Europe
Mar 30th
Festo are once more impressing everybody with the unveiling of their new SmartBird (or robotic seagull) technology (see also their incredible Elephant-inspired robot arm and AirJelly). The new flying robot which is inspired by the herring gull was shown autonomously flying both indoors and outdoors.
This is what the company has to say about their creation,
SmartBird is an ultralight but powerful flight model with excellent aerodynamic qualities and extreme agility. With SmartBird, Festo has succeeded in deciphering the flight of birds. This bionic technology-
bearer, which is inspired by the herring gull, can start, fly and land autonomously – with no additional drive mechanism. Its wings not only beat up and down, but also twist at specific angles. This is made possible by an active articulated torsional drive, which in conjunction with a complex control system makes for unprecedented efficiency in flight operation. Festo has thus succeeded for the first time in realizing an energy-efficient technical adaptation of the natural model.
The engineering team behind this magnificent flying machine has achieved an incredible task minimizing the weight of the robotic bird while maximizing lift from its flapping winds. The robotic seagull is one meter long, weighs less than half a kilogram and boasts a 2-meter wingspan. To compare with the real bird, the European herring gull is a bit over half a meter long, has a wingspan up to 1.5 meters and weighs around one kilogram.
Festo have designed SmartBird as a platform for developing and testing new ideas in lightweight construction and aerodynamics that might transfer to the development and optimization of hybrid drive technologies and open the gates for new advances in automation. However, considering how realistic this robotic bird looks like, I wouldn’t be surprised that in a a few years these are not used by the military or law enforcement agencies for surveillance; but to get to that point, SmartBird will probably need to improve its range of autonomous operation and carry a sensor pack for data gathering both of which seem to be lacking in the current model.
At any rate, words can’t do this robot justice so enjoy the below video.
[source]
Google’s robotic cars unveiled
Oct 10th
Out of the blue and in a post titled “What we are driving at” written by Stanford professor Sebastian Thrun of Grand and Urban Challenge fame (at least to the public because he is otherwise very well known in research circles), it was unveiled yesterday that Google has been developing robotic cars for urban environments. And they have been testing these autonomous vehicles in our cities.
Our automated cars, manned by trained operators, just drove from our Mountain View campus to our Santa Monica office and on to Hollywood Boulevard. They’ve driven down Lombard Street, crossed the Golden Gate bridge, navigated the Pacific Coast Highway, and even made it all the way around Lake Tahoe. All in all, our self-driving cars have logged over 140,000 miles. We think this is a first in robotics research.
The self-driving cars come equipped with laser, radar and vision sensors much like the cars that competed in the Urban Challenge a few years ago. Google has automated a handful of Priuses and an Audi TT as part of this project. Stanford is also preparing an Audi TTS for autonomously driving to the top of Pikes Peak.
I started this article by saying that the announcement came out of the blue but we actually suggested that this project was in the works back in 2007 when Google licensed some of Stanford’s technology used in DARPA’s competitions and also hired professor Thrun who in the past was rumored to be working on his own technology start up with the aim of mapping cities.
So what should we expect the outcome of such a project to be? First of all, expect near real-time updates of Google maps. Second, expect that these technologies will eventually become available for all cars which will drastically change the way we commute using our favorite means of transportation. Autonomous cars promise to eliminate road congestion allowing more cars to share the road by driving closer together. But before any of this becomes a reality, it has to be shown that the robotic vehicles are safe especially in the early days when these marvels of technology will have to share the road with human drivers.
[source]
NASA’s Curiosity rover set for Mars
Sep 22nd
NASA’s new rover that is set to be deployed to our nearest planetary neighbor in 2012 has very big shoes to fill after the outstanding success of its predecessors. But “Curiosity” seems to be equipped with many gadgets to allow it similar success. Scientists have described it as the “most advanced vehicle sent to a planet yet”, and have great expectations that it can find some interesting things up there.
On one hand, the design largely follows the concept of the Spirit Rover, as why fix something that ain’t broke? It moves on a six-wheel system, but it is also a bit larger than Spirit, giving it extra power and mobility. Each wheel has its own drive motor, and the vehicle can perform a full 360-degree turn rather easily. Scientists can’t account for all the slumps and bumps on Mars, but Curiosity will hopefully be able to handle the majority of them.
As suggested by a $2.3 billion price-tag, it also comes with the latest high-tech available. Perhaps its coolest feature is its laser rock-vaporizer, which is exactly what it sounds like – a powerful beam that can remove obstacles in its way although it is most likely going to be used for more benign science experiments. Not forgetting the exploration part, Curiosity also will employ various sample-collecting techniques with which it will drill into the Martian soil. It’s 3D cameras will also be useful when mapping the topography of the area, giving us the most detailed view of the red planet yet.
As with previous rovers, Curiosity will search for clues of life on Mars – not little green men ,of course, but small building blocks on a molecular level that may tell us a lot about the history of our neighbor. All this is in preparation for sending a human there, which hopefully will happen sometime in our lifetimes.
[ NASA ]
The HRP-4 humanoid robot unveiled
Sep 16th
Kawada Industries and the National Institute of Advanced Industrial Science and Technology (AIST) have unveiled their new humanoid robot HRP-4 which is an upgrade on their older robots. The new humanoid stands 151cm tall and is much thinner than its predecessors. For Japan, HRP-4 is another step forward towards creating useful mechanical worker to combat a forthcoming shortage in the labor force and care for a fast aging population.
Other than terrestrial uses, let me remind you that the Japanese also plan on developing humanoid robots for exploring the Moon.
Going back to HRP-4, we should note that it is not a machine developed just for nice demos. Its creators are serious about making it available commercially to universities and other research institutes that can afford its $300,000 price tag. According to the recent press release, HRP-4 will go on sale as early as January 2011. I should mention that the HRP-4 will sell for $100,000 less than Willow Garage’s PR2. However, the price difference probably reflects that other than the obvious difference between the two robot, i.e., legs versus wheels, there are many differences in the number and types of sensors, overall hardware, and software.
Time will tell if the humanoid HRP-4 will be a success or not. In the meantime, we can all marvel at its slim design and skills in the video below.
Willow Garage puts PR2 up for sale
Sep 12th
Good news everybody (you are meant to read this in professor Farnsworth’s voice), after a successful beta program that placed 11 robots in the hands of university researchers at no cost to them (but at a cost of $4.4 million to the company over a period of 2 years – recipients), Willow Garage recently announced that the Personal Robot 2 (PR2) is now available for sale.
The bad news are that it is a bit too expensive and only those labs with $400,000 to spare will be able to afford one. However, Willow Garage is willing to recognize major open source contributors offering them an incredible $120,000 discount. Also, keep in mind that millions of dollars have been spent in the design and development for both PR2′s hardware and software so from that point of view the robot is actually a bargain purchase for most universities (watch the PR2 folding towels).
The PR2 comes with 2 arms and grippers, the omnidirectional base and several sensors including cameras and LIDAR. On-board processing power is provided by 2 Xeon processors with 8 cores and 24GB RAM. Of course, the robot comes equipped with Gigabit Ethernet and Wireless communication options for access to even more compute power off-board.
It is hard to deny the fact that PR2 is by far the most successful open source robot effort ever attempted (learn how the PR2 works). Congratulations to the team and many wishes for much more groundbreaking work in the near future.
Brain-controlled prosthetic limb most advanced yet
Aug 19th
Scientists at the Johns Hopkins University Applied Physics Laboratory (APL) were awarded no less than $34.5 million by the Defense Advanced Research Projects Agency (DARPA) to continue their outstanding work in the field of prosthetic limb testing, which has seen them come up with the most advanced model yet.
Their Modular Prosthetic Limb (MPL) system is just about ready to be tested on human subjects, as it has proved successful with monkeys. Basically, the prosthetic arm is controlled by the brain through micro-arrays that are implanted (gently) in the head. They record brain signals and send the commands to the computer software that controls the arm. To be honest, it will be interesting to see just how these hair-chips are attached to the brain, but the APL say clinical tests have shown the devices to be entirely harmless. The monkeys didn’t mind them too much, at least.
MPL joins other efforts to create artificial limbs including prosthetic fingers from Touch Bionics and the also DARPA sponsored Luke Arm.
The robotic arm itself weighs nine pounds, which is about as much as a real limb, and provides just as much dexterity too. Besides tasks like moving each individual finger and rotating the wrist, it is capable of 22 degrees of freedom, and reacts with speed and agility to the user’s commands. The MPL is a vast improvement on previous models, and can allow patients a level of freedom they never thought they’d have again. Initially, the design will be used on people with spinal-cord injuries, who have lost nearly all movement and would benefit the most from using the robotic limb.
This is all permitting the human-subject tests do go well, but there is no reason why they should not. The revolutionary approach may lead to similar innovations for other limbs, which would help people with all kinds of injuries and diseases lead a much more independent and normal life.
[source]
NAO robot to help autistic children
Aug 17th
We know that the Nao robot is the latest innovation in the field of human-robotic relations. It is a small robot capable of interacting and understating us on an emotional level. Fronted by Lola Canamero, a computer scientist at the University of Hertfordshire, the project is designed to help us get more used to seeing robots in our everyday lives.
The “emotional bond” that we can establish with the knee-high Nao is made possible by his profile-recognition system. By using cameras that can sense and understand changes to the human face, and remember past interactions with specific individuals, Nao knows who he is talking to, and in what mood that person is. Accordingly, the little robot can use simple facial and body gestures to react to our emotions, and can show happiness, worry, excitement, and even anger and fear.
Creating even the simplest of robots that can get angry at us doesn’t sound like the best idea, but then again it wouldn’t be correct to say it can mimic human emotions if we leave out the negative traits. Nao learns to understand us better encounter after encounter, and possesses the development skills of a growing 1 year-old child. In fact, it is modeled after the behavioral and cognitive patterns of human and chimpanzee babies.
Nao is certainly an impressive robot, but it is it’s simplicity and straight-forwardness that allows it to help autistic kids. These children often have learning problems and have a hard time expressing and reflecting emotions, which is Nao’s area of expertise. Curiously enough, the kids have shown to be more comfortable with the robot than with other people, because they have more control in the situation. The goal will be to stimulate their emotional responses and improve their interactions with other people.
[source]
MIT Researchers develop new image-recognition software
Aug 12th
One part of the human brain that has not been focused on too much has been its visual-perception ability, which allows it to recognize, differentiate and quantify different objects. This is, of course, a vital aspect of the brain’s function, which directly links what we see through our eyes to the way we actually understand things in our mind. Now, scientists at MIT’s McGovern Institute for Brain Research have developed software that studies and learns from this ability, gathering information that can pave a new way forward for artificial intelligence.
The software, based on a mathematical model, studied a number of test subjects, who were given different object-recognition tasks. In one experiment, a group of people were asked to look at a picture of a street, and count the number of cars and pedestrians on it. The software’s eye-tracking system recorded the way their eyes moved, and then began predicting where they would first look to next in other tests. The information gathered confirmed that the human brain creates an outline of the image the eyes see, and initially recognizes objects that are more significant or stand out in some way. In a related set of tests, the people were given two objects – a square and a star, and their attention was split on both equally. When given a whole bunch of stars but only one square, the eyes first focused on the square, or the more unique shape. This whole process happens in a split second, but it shows that we always look for something special or unique in whatever we encounter.
When testing out the software on its own, it created a similar spatial map on which it recognized different objects. To do so, it ran down a list of features specific to the object it was asked to identify, and only focused on those that correspond to the said object, while ignored everything else that did not match the description. To truly mimic the human brain, this process will also have to become much faster, but MIT’s researchers are confident they can get it close to that level.
[source]
ROCR robot swings up walls
Aug 10th
What do you get when you combine a monkey and a grandfather clock? The ROCO–Oscillating Climbing Robot, that’s what. Designed by researchers at the University of Utah, this 1.2 pound robot that measures only 12.2 X 18 inches can climb up all kinds of walls, even carpeted ones, with impressive efficiency.
While it does not actually look anything like a monkey, or a clock, it copies their movements to produce its unique style of climbing. It does so by swaying back and forth with its large pendulum tail, which gives its momentum and allows it to inch up swing by swing. It’s hook-like claws allow it to attach itself to the surface without doing any damage, which gives it stability and prepares for the next motion up. This stop-start climbing method is what monkeys, particularly gibbons, use to jump up on trees.
Recent tests showed ROCO climb up an 8-feet tall carpet wall in about 15 seconds, which is rather impressive. It moves up with 6.2 inches per second, but actually the most significant achievement was that it reached a 20% efficiency rate. This is only 5% less than a car engine. In previous experiments, the researchers said their biggest problem was making sure ROCO did not fall of the walls it was climbing, so this is a big improvement.
One of the assistant professors who worked on the project, William Provancher, has said that in the future, these robots can be used for climbing up and expecting tall buildings, bridges, nuclear facilities, and all kinds of other places where a human worker would be in danger. To accomplish such tasks, the robots will have to bulk-up in size, because right now it looks like a strong wind might be all that’s needed to knock them down.
Watch a video demonstration of the Utah robot climbing up a wall.
Recent Comments