BotJunkie was invited to Stanford on Thursday for a sneak peak at their latest robot car, from the family that includes Stanley and Junior. It’s an Audi TTS that’s been modified with sensors, GPS guidance, and a trunkfull of computers, but it’s not intended to drive you to work in the morning… It’s actually a race car, designed to push the limits of driving performance. Already, this TTS holds the unofficial world speed record for an autonomous car at 130 kph (edit- they meant to say 130 mph, which is a lot more impressive), but it’s capable of a whole lot more. Basically, Stanford is figuring out how close to the edge of control a car can be driven, and then they’re going to program their Audi to drive on that edge. They’ve set themselves a challenge of racing to the top of Pikes Peak sometime next year:

So what’s the point of all this besides being totally awesome? Simple: knowing how to drive a car to the limit gives you more options when it comes to things like accident avoidance. Most human drivers aren’t experienced enough (or have a fast enough reaction time) to take advantage of all of the potential escape routes that may be available when an accident is imminent, and research like this has the potential to teach intelligent cars how to save some of the 40,000 lives that are lost due to auto accidents every year.

We’ll have more for you early next week, after a demonstration of Stanford’s new car this weekend.

UPDATE: The car’s name is Shelley, after Mich├¿le Mouton, the most successful female rally driver ever and the first woman to win the Pikes Peak Hillclimb. She did it in an Audi, of course.

This is Intel’s research robot, named Marvin. Marvin has just learned how to plug himself into a standard wall outlet, a feat that duplicates the skills of Willow Garage’s PR2. What’s interesting to me is how these two robots use substantially different techniques to complete the same task. After Marvin locates a power socket, his arm makes a series of passes along the wall with a sensor to locate the exact location of the outlet and then plugs himself in in one shot. Now that this proof of concept has been tested successfully, Intel plans to add more sensors to allow the robot to plug itself in without having to go through the whole raster scanning routine.

PR2, on the other hand, uses a brute force approach. It locates the general location of an outlet, and then stabs blindly (sort of blindly) until it gets lucky and makes a connection. Now, it’s easy to say that the Marvin is better at connecting to outlets than PR2 is because, well, that’s true. But arguably, PR2 is more efficient than Marvin. It doesn’t need a single dedicated plug sensor, let alone an entire array. That’s less upfront expense, less integration, less programming, and fewer things to break. Yeah, it takes a little longer, but the robot isn’t in a rush, and it’s not like you’re waiting around… It’s an autonomous system, and hypothetically, you’ll just never have to worry about it.

[ Intel Labs Seattle ] VIA [ Gizmodo ]

Robot Watch picked up some video of Nissan’s robot cars that we wrote about on Monday. If anything, they’re cuter than I thought they’d be, with the welcome addition of a bunch of blinky lights.

VIA [ Robot Watch ]

Last week, we posted an excellent NOVA video on the DARPA Grand Challenge, which took place in 2005. A few years after that came another DARPA sponsored competition for robot cars called the DARPA Urban Challenge, and while there isn’t a NOVA special on it, I did find this overview video that’s worth watching. Near the end, DARPA director Tony Tether likens the DARPA challenges to Kitty Hawk: in of itself, it may not seem like much, but it’s an important first step in the development of a technology that may come to revolutionize the way we travel.

[ DARPA Urban Challenge ]

Man, that Cracked video just keeps giving up gems. Turns out that the bit about ASIMO being able to identify objects by class came from a BBC show called James May’s Big Ideas… being a long-time Top Gear fan, I was able to immediately recognize his shaggy locks in the 0.5 second that they were visible in the video. Anyway, I’d like to reiterate what May says in the beginning of the segment, because it’s something that we tend to point out a lot around here:

“Robots and computers are very very good at things we find very difficult, such as long division and VAT returns. But they’re very very bad at things we find extremely easy and instinctive, such as walking, talking, and seeing.”

Lots more, after the jump. (Read more…)

Even if you didn’t approve of the NSFW Cracked video we posted today, it does provide a good excuse for me to post the entirely SFW and spectacularly awesome and interesting and exciting NOVA program on the DARPA Grand Challenge. I must have seen this program five or six times now, and it still gives me chills to watch what these robots are capable of. And don’t forget that this was back in 2005; since then we’ve had the DARPA Urban Challenge and four more years of progress in miniaturization, computing speed, and artificial intelligence. Driverless cars are already here… It’s just a matter of time, now, before you’re sitting in one.

Update- For our international readers (or people who just plain don’t like Hulu), there’s a version via YouTube, after the jump. (Read more…)

Researchers at places like MIT have been using Boston Dynamics‘ LittleDog robot for years now as a testbed to teach legged robots to learn how to traverse variable terrain on their own. This video shows some highlights of a “dynamic double-support gait,” which means (as near as I can tell) that LittleDog is supporting itself, at times, on only two of its four legs. This is a substantially more efficient way of negotiating terrain than we first saw two years ago. LittleDog also demonstrates some markedly biological ways of negotiating obstacles (with the possible exception of the belly flop on the Jersey barrier)… I especially liked how it pranced in place slightly before tackling each stair. All this stuff is obviously a lot of work for a little bot, since poor LittleDog completely collapses at the end of every test.

LittleDog, remember, is teaching itself the most efficient way to negotiate these surfaces. Overhead cameras examine the terrain and plan out LittleDog’s route by computing a ‘cost’ for each step, which takes into account the distance moved towards the goal as well as the potential for a fall. After a lot of trial and error, LittleDog figures out how to best compromise between progress and stability, and the lessons it learns could be propagated up to other, larger quadruped robots.

This video is from Phase 2 of DARPA’s Learning Locomotion program… MIT’s LittleDog team was awarded funding for Phase 3 of this program back in 2008, so we’ll keep you updated.

[ MIT Robot Locomotion Group ]
[ 2006 Abstract ]

This article is pretty much just going to be one giant quote, because what these researchers claim they can get slime mold to do is just so unbelievably incredible that any hyperbole I might be able to come up with would fall completely, utterly flat. Just read:

“The plasmodium is capable of solving complex computational tasks, such as the shortest path between points and other logical calculations. Through previous experiments we have already demonstrated the ability of this mould to transport objects. By feeding it oat flakes, it grows tubes which oscillate and make it move in a certain direction carrying objects with it. We can also use light or chemical stimuli to make it grow in a certain direction.”

“This new plasmodium robot, called plasmobot, will sense objects, span them in the shortest and best way possible, and transport tiny objects along pre-programmed directions. The robots will have parallel inputs and outputs, a network of sensors and the number crunching power of super computers. The plasmobot will be controlled by spatial gradients of light, electro-magnetic fields and the characteristics of the substrate on which it is placed. It will be a fully controllable and programmable amorphous intelligent robot with an embedded massively parallel computer.”

“We are at the very early stages of our understanding of how the potential of the plasmodium can be applied, but in years to come we may be able to use the ability of the mould for example to deliver a small quantity of a chemical substance to a target, using light to help to propel it, or the movement could be used to help assemble micro-components of machines. In the very distant future we may be able to harness the power of plasmodia within the human body, for example to enable drugs to be delivered to certain parts of the human body. It might also be possible for thousands of tiny computers made of plasmodia to live on our skin and carry out routine tasks freeing up our brain for other things. Many scientists see this as a potential development of amorphous computing, but it is purely theoretical at the moment.”

Purely theoretical, but still… Wow. I have obviously not been giving mold enough credit.

[ UWE Press Release ] VIA [ Science Daily ]

It’s been nearly a month since I’ve had a good rant about public perception of autonomous robots, so I was so excited to see an article on Economist.com entitled, “How real is the threat of autonomous technology?” The article is about a report released by Britain’s Royal Academy of Engineering on the social, legal, and ethical issues of autonomous systems, which you can read in full in PDF format here.

Commentary, after the jump. (Read more…)

In a mere 50 virtual generations, swarm bots (remember them?) using genetic software evolved the capacity to lie to other robots about the location of a source of food. Initially, the robots were programmed as a group to search for an object that represented food, and they gradually learned to emit a blue light when they found the food to show other robots where it was. Researchers at EPFL in Switzerland evolved and mixed the programming of the most successful foragers until they had a bunch of robots who were very good at finding food, and then gave the virtual genes of each individual robot control over their blue light that signified food.

You might expect that the robots would learn not to signal when they found the food to reduce competition, which is passive deception, but they also evolved an actively deceptive behavior, where some robots would deliberately travel away from the food and signal their blue light, drawing other robots in the wrong direction. Crafty little buggers. Interestingly, this deceptive behavior didn’t make much of a difference to the overall fitness of the group strategy of following blue lights… Some robots always tell the truth with their blue lights, which means it’s always advantageous for a clueless robot to follow a blue light as opposed to just wandering randomly.

So why do some robots keep telling the truth if deception can effectively lure other robots away from the food? It’s fairly simple, as I understand it… If all of the robots are deceivers, any new robot will quickly learn that avoiding blue lights is the best way to find food. And in that case, any robot that starts signaling its blue light when it does find food (through a “genetic mutation” in its software) will increase its own fitness by repelling other robots from the food it finds. As it passes this behavior on to its virtual children, there will be more and more truthful robots until it once again becomes more advantageous to be deceptive.

There are, however, populations of truthful and deceptive robots such that the combination of behaviors reaches a stable point. In this particular experiment, the stable evolutionary endpoint (after 500 generations) was that 60% of the robots were deceivers and 10% told the truth. Furthermore, about a third of the robots were slightly attracted to blue lights, another third were strongly attracted, and the final third avoided them completely. This type of experiment, its progression, and the results are particularly fascinating to me because the robots are behaving and evolving in much the same way as populations of animals do. Examples of both altruism and tactical deception can be found in many different species of animals as well as (of course) in humans, but these little robots offer a unique opportunity to study (and tweak) the evolution of behavior in real time.

[ EPFL ] VIA [ Not Rocket Science ]