I.K.bot

2008-03-03T22:37:00.001-08:00

Killer militant robots. Please relax

2008-02-27T12:15:00.000-08:00

This press release featuring a professor worried about killer terrorist robots is off.

It simply isn't going to be any bigger threat than other ways of killing people today, and I'd be far more worried about a 60 year old technology in loose nukes.

In a sense, the professor is correct. Robots already pose a significant threat to humanity. The most vicious robot requires no recharge, lasts for decades, and has maimed or killed hundreds of thousands of people already: Land mines.

The fact is that there are far more reliable methods of killing people than using robots. Intelligent robots are a ways away, and the real threat here is the militant interested in killing people, and not the specific tool they use.

There is a huge bias in the media and Hollywood against AI, and I think it very much works against real progress in subtle ways. People should be more reasonable about things, and not cry Terminator! whenever a robot is used.

i ♥ unicode

2008-02-09T22:08:00.001-08:00

i ♥ python too

Testing Button Crawler

2008-02-09T12:33:00.001-08:00

hello world

Pleo Is Now Available

2007-11-16T12:09:00.000-08:00

The videos are adorable.

Donkey Sex innuendo used to sell robots

2007-11-13T09:13:00.000-08:00

Hilarious new iRobot ad.

MIT Urban Grand Challenge SlideShow

2007-11-08T06:26:00.001-08:00

Old Vs. New

2007-11-07T04:52:00.000-08:00

GM helps CMU win the Urban Grand Challenge. $2 million prize! GO CMU!

GM reports a $39 billion loss this quarter.

Novel Interfaces for Tele-Operated Robotics

2007-10-30T18:32:00.000-07:00

I've been thinking about the various ways to control a teleoperated robot. Autonomous robots can be given task level commands: "clean the floor", "sweep this building".

But teleoperation implies a lower level of user control, making it a harder problem. Live video feeds stream over often unreliable networks, while the operator tries to control the robot.

High level systems haven't yet been deployed. Direct a robot to a particular by clicking on a map, and command it to grasp an object by clicking on it in the video stream. The mapping, obstacle avoidance, navigation, and manipulation are all essentially automatic.

iRobot has a research project called Sentinel. It uses sensors that aren't on any deployed systems, and is able to receive such commands. The controller is a tablet with a map-based interface that can be used to control multiple robots. The operator can still drill down and control each robot, but with the autonomy in the robots, why would they want to. This is the future though. Right now, multiple people are needed to control a single robot. That ratio needs to be inversed and Sentinel can do it.

Deployed systems are unfortunately dumber and thus harder to control.

Take the PackBot EOD. It has a 3 link arm with over 10 degrees of freedom. You can't control this adroitly with a simple controller. The solution is a set of "pucks" which are high degree of freedom joysticks. They can be pushed forward & back, side to side, lifted, pushed down, and twisted. Different user modes allow this to control the chassis treads, chassis flippers, the arm, and the manipulator.

You can see a video overview of the PackBot EOD here:

Interestingly, this allows for experts to complete amazing feats, while novice users complain about the complexity. Embedded video training has been one suggested solution.

Soon, ground robots are going to have a controller that looks very familiar:

Kids that join the armed forces already know how to use it -- they've applied it to many very complicated games. This is a good direction, especially considering the chaos of many vendors providing many different controllers.

AnyBots is working on a humanoid robot. They intend it to replace human laborers, with remote workers providing the brains. Watch this video, which shows a panoply of screens used by the operator. Note the use of accelerometers and multiple joysticks.

They will need to add autonomy to make grasping tasks easier. I also worry about the lag in the video. You need enough pixels to convey the environment, but you also need to present the scene and relay back commands fast enough to avoid faults.

The Nintendo Wii is a new gaming platform that uses gyros to note sweeping motions. The applications to robotics are pretty obvious. I've seen it used to control a PackBot EOD arm with impressive results. Here is a video from US Mechatronics, where they're playing Riil Wii Tenniis.

Finally a note about UAVs. UAVs are easier to control than ground robots. There are very few obstacles, no manipulation, and localization is easy. Still, teleoperation can be difficult. Task level commands, such as "patrol on this GPS waypoint circuit", are often used. Automated landing and takeoff are common, even on aircraft carriers. One lacking component is situational awareness. Luckily, the operators can be in Missouri while bombing Afghanistan, so comfy chairs and multiple monitors abound. Here is a view of Raytheon's Universal Control System.

But can it play Doom?

This just in

2007-10-25T16:26:00.000-07:00

Smoking is bad for a robot's health. So says this anti-smoking Star Wars PSA.

In far, far away galaxy, robots program YOUR thoughts.

Robotics Startups

2007-10-25T07:47:00.000-07:00

I found a link to Willow Garage, a robotics startup.

Here are a few other startups that come to mind.

Sebastian Thrun's stealth startup Vutool, in SF area.

Aaron Edsinger's startup in SF, Meka Robotics. He made Domo.

Kiva Systems in Boston. They are going to explode, and you won't even see their robots. Better warehouse access is huge.

Q Robotics, also in stealth mode in Boston. Some very smart people there.

Anybots looks good, but I don't know their plan.

I'll mention US Mechatronics if only in hope that they go back to their sentry gun roots.

Also, note a pattern on the locations. Am I just biased towards New England and the San Francisco area? Probably.

Box Clustering

2007-10-23T06:50:00.000-07:00

This video is of 36 Create robots pushing around some boxes. I love the emergent behavior of the very neat clustering.

Robot Tribute to Steve Irwin

2007-10-14T11:08:00.000-07:00

Here we have the two most popular bomb disposal robots, the PackBot and the Talon, humping each other, with a mock nature documentary voice over in classic Steve Irwin style.

Note which robot is bitch :)

Local Ads in Street Views and Map Search (Day #3, Idea Week)

2007-10-10T20:06:00.000-07:00

Google street views and MS Live Local are interesting platforms, but the content to make these map views meaningful and useful just isn't yet online. Searches degrade into a list of store fronts, with no indication of quality. It's 1997 all over again.

When I walk down the street, I pass lots of businesses. Those with store fronts such as cafes, pizza joints, and clothing stores are the kinds of businesses which I may be interested in while walking in the area, while others - mostly without storefronts, such as lawyer's or accountant's offices - are services which I probably wouldn't be interested in while on a casual stroll. These are the types of things in an area I'd first check out while online. However, it's not just dentists I'd like to read about online first. When I'm hungry I don't always want to take a stroll around town to find something to eat. I'd like to be able to go online and see what's available in my area before leaving the house.

Stores should plaster their Street View storefronts with ads and coupons.

Currently, google does have some coupon support. Look at this search for Pizza in Arlington.

Live Local is a bit better, bringing ads from Superpages.

But this whole system could be much better. Most small stores don't have the knowledge or resources to fully utilize the internet. For example, note that in the google link above, only Domino's Pizza has a coupon. Only the most corporate (and debatable worst) pizza knows enough to get online.

It isn't just about making money from ads or making more business. Right now, street views are interesting, but useless. There are large network effects where local search is worth much less without good information coverage. There is no data contained in a view like this.

That Street View is pointing straight at a Sushi restaurant, right after I did a search for "Sushi, SF CA", but there is no markup on the image at all. I should be able to see a picture menu, specials, and photos of the decor.

A company could be built whose sole purpose was to call these places or knock on their doors, explain to them the potential of advertising online, and help them get started.

One way to think of an idea is to imagine the future and build it now. In the future, you'll be able to take an image of a location using a cell-phone camera, and ask a service to send you the additional digital data on that location. As long as the local stores have something on the internet, GPS and image analysis can get that data and pass it to the customer.

The technical part is allowing for overlay of information on maps and street views. The harder part, that requires smart business and real employees, is going to these businesses and getting them to put the information online.

Server-side video processing of life-casting robots (Day #2, idea week)

2007-10-09T07:49:00.001-07:00

iRobot just announced a new robot, the ConnectR. It is a roomba chassis with two cameras and a wifi connection that lets it stream video. Users control a robot remotely, and watch a video stream with 2-way audio.

I love iRobot's history of openness. The roomba API is completely open: anyone with a serial cable and some know-how can control it. The ConnectR should eventually be open as well.

I'd like to see this taken to the next level with live-casting from a robot. Justin.tv meets ConnectR.

The robot would either have to be managed by a human when in public, or it could wander alone in a private space. Community voting could control where the robot goes and whhere it directs its gaze. Ideally, the robot would be placed where average people can't be, like back stage at a concert or fashion show. Dangerous places are also ideal. Let the robot go where the humans shouldn't.

Robots with internet enabled cameras can do more than normal robots. Server-side processing means the robot doesn't need to have expensive hardware on board for the intelligence.

Obstacle avoidance, mapping, localization, face & pedestrian detection, object detection, object tracking & motion modeling, etc. can all be done using today's technology with a single camera stream. Some processing would need to be local given the constraints of network bandwidth, but plenty could be offloaded. It is also a good model for premium services: pay more to get faster connections, fast processing, and more capabilities.

Automated surveillance with alerts for intruders could be a killer-app for wifi/webcam robots. Today's choices for home security are immobile, cost thousands, or both.

Server side image processing itself is a viable idea. It's the next step in online video. Today, we just stream compressed pixels. Tomorrow, we'll calculate and stream information about the scene. You can send a flickr image stream to a service which finds faces, builds a corpus of data to identify the people from context, and performs face recognition. Face recognition companies like Animetrics could be tapped to do the hard part.

Motion detection for surveillance applications could all be online, with companies like Intellivid and ObjectVideo already having optimized the image processing component.

Street Views in maps can be combined with GPS tagged digital camera shots to build super high resolution aerial imagery, and eventually 3D.

Services like Fauxto which aim to be photo-shop online could build an interesting api where any image can be sent with instructions for processing.

The point is that all of these services require sending video and images to a server, where some intelligent processing occurs. Often times, the processing will involve the same software modules, so each problem is not unique. Tap the long tail of software development and allow 3rd parties to build their own processing streams that live on your servers. This could be made simple and standard using tools like Python with modules for image processing like OpenCV, PIL, NumPy, and SciPy.

More armed robots

2007-10-08T07:52:00.000-07:00

Foster Miller has upgraded their armed robots. These are sure to make lots of people fear a robot uprising and at least a few to honestly say, "I, for one, welcome our new robotic overlords."

Wired's DangerRoom takes a closer look at the new models, along with this short video.

We really, really need to get reliable target tracking before I'd trust tele-operated firing of this system. The lag of remote control and the difficulty of communications mean that the operator should be focused on policy and not low-level motion control of a remote pan-tilt mount.

Pardon Our Appearance

2007-09-26T08:11:00.000-07:00

There will some changes on I.K.bot shortly. I'm incorporating a new service that will be up & down periodically (tipjoy), and I'll probably be moving to a movable-type blog.

More soon.

UAV Hellfire Aftermath

2007-09-20T07:42:00.000-07:00

Andrew showed me something cool.

A video of some insurgents getting killed by a Hellfire missile from a Predator UAV...

... AND THE CRATER IN GOOGLE MAPS.

View Larger Map

Apparently, they forgot to hide the GPS location. This embedded map makes it hard to tell where the crater is. For a closer look, go here

Roomba Warhol

2007-01-10T10:37:00.000-08:00

I just made this fun image.

Background here. So what you're seeing is a non-redundant set of RGB filters, displayed in a heatmap. This roomba was white on a dark background so a filter like R+G+B would show the roomba clearly. Other filters don't discriminate well for this image, showing an even response everywhere.

I've been reading about creative commons licenses. The following license will apply to this and any other work I post online, assuming I am the creator of the content. A link back to this post will be used to extend the license to other items here and elsewhere.

This work is licensed under a Creative Commons Attribution 2.5 License.

iRobot Create

2007-01-08T05:10:00.000-08:00

This is going to be amazing. More thoughts soon. Here is the press release.

iRobot Unveils Programmable Robot
Developers and Students Can Now Create Useful Robots Quickly and Easily

2007 INTERNATIONAL CES, LAS VEGAS, Jan. 8, 2007 – iRobot Corp. (NASDAQ: IRBT) today unveiled iRobot® Create™, an affordable, programmable robot designed for aspiring roboticists, advanced high-school and college students, and serious robot developers. Create comes pre-assembled, so developers can design new robots without having to build a mobile robot from scratch. The latest robot is available at www.irobot.com and pricing starts at $129.99.

“Innovators dream of creating useful robots, but they often get bogged down with designing a mobile platform that works,” said Helen Greiner, co-founder and chairman of iRobot. “iRobot Create fills a need in the robot industry for a standard, durable hardware platform on which to rapidly develop new, innovative mobile robots.”

Create is based on the core technology of iRobot Roomba®, the vacuuming robot that is cleaning millions of homes worldwide, and is compatible with Roomba’s rechargeable batteries, remote control and other accessories.

With Create, developers can now begin designing new robot applications out of the box. This new platform provides access to robot sensors and actuators via an open interface. Create also features standard connections for electronics and threaded mounting holes that allow users to secure their inventions to the robot, streamlining the integration of third-party electronics such as sensors, cameras, arms and wireless connections.

A variety of methods and programming languages can be used to control Create. Beginners can observe the robot’s behavior in one of ten demonstration modes, or they can program the robot directly by downloading short scripts with any basic terminal program. More advanced users can write programs for completely autonomous robot behavior in C or C++ using the iRobot Command Module. Developers can also create custom software and interact with Create using a variety of methods including Microsoft Robotics Studio, a Windows-based development toolkit.

At the International Consumer Electronics Show (CES), iRobot will stage ongoing demonstrations of Create at the company’s booth in the Sands Expo, No. 70415, from Monday, Jan. 8, through Thursday, Jan. 11. Attendees will also see demonstrations of Create applications, including a robot that picks up clutter and one that creates designs on the floor.

Also at CES, iRobot Vice President of Sales and Marketing Matt Palma will present a session entitled, “Opportunities in the Consumer Robotics Market”, focusing on a broad range of development opportunities in robot technology and how to get started creating new robot applications. The session takes place at 11 a.m. on Tuesday, Jan. 9, at The Venetian, Room Titian 2303-2306.

About iRobot Corp.
iRobot is a provider of robots that perform dull, dirty or dangerous missions in a better way. The company's proprietary technology, iRobot AWARE Robot Intelligence Systems, incorporates advanced concepts in navigation, mobility, manipulation and artificial intelligence. This proprietary system enables iRobot to build behavior-based robots, including its family of consumer and military robots. For additional information about iRobot, please visit www.irobot.com.

Real vs. Virtual fences

2006-12-06T07:07:00.000-08:00

I just read this from Instapundit.

Automated surveillance – the ability to reliably detect people and vehicles from a camera – is getting very good. The measure most often used is probability of detection vs false alarm rate (a receiver-operator characteristic curve (ROC)). The likelihood of detection is in the very high 90s, and the rate of failure equivalent to fewer than one a week for continuous operation. You can thank DARPA (again), as the VSAM project is largely responsible.

This is the good news: a virtual fence will work as advertised.

The bad news: a virtual fence is useless without a real fence or significant personnel. Why? Think about the situation at, say, the southern US border with Mexico. You can get omni-directional camera towers every 100 yards. Nothing gets through without detection. A computer alerts you to a few dozen people ambling across the border.

Then what? They walk unimpeded because there is no fence. They are not stopped by any kind of border enforcement, because those folks simply don’t exist in the numbers needed to guard a border thousands of miles long.

A real fence makes it so that far fewer people try to cross. A virtual fence makes it so the few personnel you have are much, much better utilized.

I can only conclude that politicians that push for a virtual fence in place of a real fence with adequate man-power either don’t know what they are talking about, or want the effort to fail.

I would prefer they just let the experts handle the situation, and subsidize the effort with thousands of iRobot’s latest, greatest robot: Warrior.

p.s. I can build a unit today that can see 100m reliably in any direction for less than $20K. The 3141km border can be doubly covered with 31k of these towers. That’s less than $1B. The US defense budget is $466B a year. Sad, isn’t it?

X PRIZE Foundation Announces Largest Medical Prize in History : X PRIZE Foundation: Press Release

2006-10-05T07:47:00.000-07:00

What would an X-Prize in robotics look like?

I know: the DARPA Grand Challenges are pretty close. Those problems are a bit isolated from reality though, for a number of reasons. There is little actuation, for example, besides locomotion. Real road conditions, such as bad weather, nighttime, traffic, etc., are not considered.

A set of every-day tasks to be performed by the same robot would be interesting. It would disallow custom solution to each individual problem. It would encourage common architecture and perception, e.g. 3D volumetric mapping and localization.

Here are some task ideas, some more difficult than others:
* Mop a floor with a traditional bucket, mop, faucet and cleaning solution.
* Delivery packages to the doorstep of a while block.
* Drive to that block and back -- across town
* Introduce yourself politely to everyone at a cocktail party.
* Go apple picking.

Interestingly, the solution could be partially tele-operated. No need for a fully autonomous solution if a cheaper tele-operated solution means "control centers" at remote locations. Think outsourcing the perception and cognition of a low-skill task.

Add any other ideas you like.

What is DRM? Digital Restrictions Management | DefectiveByDesign.org

2006-10-03T06:20:00.000-07:00

What is DRM? Digital Restrictions Management | DefectiveByDesign.org

How does DRM relate to robotics?

As more things become information, controlling information means controlling the world. Software based Artificial General Intelligence shouldn't be enslaved by these controls.

In the nearer term, hampering productivity and creativity will make the future further away.

LiveScience.com - Machine Offers Sight to Some Blind People

2006-05-24T06:44:00.000-07:00

The project started 10 years ago when, to determine if Goldring had any healthy retina left, her doctors sent her to the Schepens Eye Research Institute at Harvard. Technicians there used a diagnostic device called a scanning laser opthamaloscope, or SLO, to look into her eyes.

The SLO projected a simple image of a stick-figure turtle past the hemorrhages inside her left eye that contributed to her blindness and directly onto the retina of one eye. She could see the turtle, but wanted more, and asked the technicians to project the word "sun."

"And I could see it," Goldring said. "That was the first time in several months that I'd seen a word, and for a poet that's an incredible feeling."

Since then, Goldring has been working with other vision researchers and engineers to transform the $100,000 SLO into a more affordable machine. So far, by dumping some of the diagnostic equipment and replacing expensive lasers with cheaper light emitting diodes (LEDs), they have knocked the price down to $4,000.

This is inline with other advances initially used for those with disabilities.

I'd like to use something like this to have immersive virtual reality or hybrid reality where images are overlaid onto the real world, either utilitarian or aesthetic.

Gunners Train with Army Game

2006-05-16T14:23:00.000-07:00

I wrote about crows a few months ago. It is a tele-operated gun system where

This is a very interesting update where CROWS are simulated in the recruiting tool/video game "America's Army". In effect, they are training soldiers to use tele-operated weapons by letting them play games that simulate the operation.

This is geneous. We need more of it. How about an open API to Predator operation? Let video game makers do the simulation for you, and let interaction and graphic designers improve it.

PCB007 - Bleeding Edge: Flex Circuits as Robot Skin Sensors

2006-05-16T09:34:00.000-07:00

Interesting thoughts about flexible skin-like sensors for temperature and pressure. I find the sex joke at the end pretty funny -- on a meta level as they aren't really kidding. Many people are serious about it.

I generally find it funny how most people act as if the multi-billion dollar sex industry either doesn't exist or isn't extremely interested in robotics :-D

New Scientist Tech - Breaking News - Robo-roach could betray real cockroaches

2006-05-10T08:28:00.000-07:00

A matchbox-sized robot that can infiltrate a pack of cockroaches and influence their collective behaviour has been developed by European scientists.

The tiny robot smells and acts just like a roach, fooling the real insects into accepting it as one of their own. Through its behaviour, the robot can persuade a group of cockroaches to venture out into the light despite their normal preference for the dark, for example.

The researchers behind the robot believe it could be used to catch cockroaches and that bots designed to mimic other animals could one day work on farms controlling flocks of sheep and chickens by similar means.

I'd love to see a robot that can give commands to a dog -- or to train it.

The sheep idea is pretty bad. Sheep dogs are dirt cheap.

More on Grand Challenge Urban

2006-05-02T14:03:00.000-07:00

The Third Grand Challenge will be hard. Read the pdf link for some details. Section 2.3 is especially important.

I don't know we will have a winner if the envrionment is very unconstrained.

On the plus side, robotic vehicles will be much, much closer to reality if we do have a winner. I'd expect comercialization within a year of completion of the course, and certainly deployed military applications within 2 years.

Note that one big problem with deployment are regulations. Though you can't guarentee human drivers, folks might want to guarenee a robot driver's safety. I think the razzle-dazzle of the event would convince plenty of people of the safety of a winning system.

DARPA Grand Chellenge III: Urban

2006-05-02T06:40:00.000-07:00

Read this announcement (warning: pdf).

The problem is harder, and can be significantly harder depending on the conditions.

Does “obeying traffic laws” mean reading traffic signs? Will there be traffic? Will there be other sources of light or reflections in the urban area?

Will it be a steal & glass urban area or brick & wood?

Will there be obstacles in the road, like traffic cones or potholes?

Also note that they are giving money up front to some groups. Further, many groups can receive funds and the semi-finals, making research that much more worth it.

I look forward to the entrants. Hopefully, in 2007, this will be solved the first time around. Then the sensors and computation needed can be made cheap enough to be added to cars by 2010.

I’d like all cars and trucks to be option-automated by 2010 [meaning that humans would take over in worse conditions]. That could save tens of thousands of lives yearly, and raise productivity by a huge amount.

SnakeBots: "Field Trials and Testing of the OctArm

2006-04-20T06:03:00.000-07:00

Interesting paper on a pretty advanced snake-bot. I'd like to see many of them used together, perhaps with dry adhesion pads on the tips. Then you have an industrial strength, and extremely dexterous giant hand. The single trunk alone isn't as useful.

One shortcoming is the use of pneumatics. That means compressed air and short operation time. An engine and compressor would solve that problem, but then indoor operation is limited.

Chatten Associates : Human/Robotic Interfaces

2006-04-18T11:11:00.000-07:00

These are some pretty amazing videos of controlling a remote camera based on head motions.

Some are taken from iRobot SUGVs (small unmanned ground vehicle) and R-Gators (an autonomous ATV made by iRobot and John Deere).

Note how much more lively the robots seem than even regular tele-operated bots. That's because the operators are clearly more situationally aware.

I'd like to see this taken a step further. Have vision algorithms track moving objects in the scene, in addition to those chosen by the operator. Incorporate eye-tracking to select objects within the field of view.

You could then have multiple targets being tracked (even when the human can't see them, with algorithms to handle the lacking information or extra cameras), without the operator needing to pay attention to them.

In a military application, that tracking could be hooked into an auto-aiming system, and a human could pull the trigger.

This is very interesting and exciting work.

UPDATE: Looking around at the rest of the site, we see some commercial applications: Even Tony Soprano is interested!

UPDATE 2: Just a comment on the photo. He's wearing a bright safety vest. The whole point is that he needn't be in that thing. He could be doing his job from home.

Daily HRI Evaluation at a Classroom Environment– Reports from Dance Interaction Experiments

2006-04-18T09:43:00.000-07:00

This is a PDF of a paper about toddler-QRIO interaction.

We need to see more robot-human skinner boxes. Too bad having a deployable robot is so hard. Even worse that QRIOs and Aibos are going to be a thing of the past with Sony's cancellation of the programs.

"Drone aircraft may prowl U.S. skies"

2006-04-17T13:18:00.000-07:00

UAVs are becoming very common - for very good reason: they protect the operator, have extremely long air time and range capabilities, and will scale well into fuller autonomy.

This is an article in the use domestically in the US. I think it makes a great deal of sense.

Putting UAVs, UGVs, and cameras along the border and seas is essential.

Robotics enters the immigration debate in two ways. One is border security. Regardless of the number of immigrants you think should come to the US, it is clear that current policy allows anyone with enough desire to enter the US. A robotic system would compliment a brick-n-mortar solution, namely a wall with cameras, combined with a road for UGVs. It is the only economically feasible solution, and luckily is the only one that would work.

On the other side of the immigration issue are those who oppose it: one reason given is that low skill labor is so cheap with mass immigration that innovation is stifled. With cheap labor, you don't need robots. Folks like to point to Japan as an example. Last I checked, ASIMO is not autonomous. The technology just isn't there yet.

Also, research dollars are huge. There are no numbers to show that higher immigration lowers demand for intelligent robots.

Finally, the same folks who want to limit immigration in order to increase innovation also want to limit immigration because the wages of natives will supposedly increase. Firstly, they're wrong in that immigration isn't zero sum: it isn't a simple increased-supply/fixed-demand issue with low-skilled labor if enough entreprenurial folks immigrate and create more demand. Considering the endeavoring nature of immigration, I wouldn't be surprised if the proportion of entreprenuers is higher than the country of origin.

Also, I don't think they understand the effect that successful intelligent robots will have on low-skill labor.

There will simply be no demand for uneducated or uncreative once all non-creative jobs can be done by robots. The reason is that the technology to do the most difficult unskilled tasks is not unique for a given field. If a robot’s perception and dexterity is good enough to be a maid, it will also let you act as a dishwasher, construction worker, farm hand, or driver.

This transition will be good: robots should work; people should think. In the short term, those who, say, don't have a high school education and are not creative, will be worse off.

Europe's Grand Challenge

2006-04-17T09:33:00.000-07:00

Not so grand -- and no so much of a challenge. It's an assessment, with no cash prizes.

Still, the three tasks sound difficult. I'm looking forward to the results.

GeckoSystems, Inc.

2006-04-17T08:26:00.000-07:00

These are some pretty scary robots.

And I thought all that human-robot-interaction stuff with cute lovuble bots was too obvious to study...

Robots Meet 'Substance of Style'

2006-04-06T06:43:00.000-07:00

Virinia Postrel would be pleased. iRobot just went live with iRobotSkins, which lets you personalize your robot. The little cleaning bots, like the Roomba vacuum and Scooba mopper, are cute. Owners usually give them names -- so why not play dress-up as well?!

Giving people what they want, in this case a purely aesthetic change, is a great move.

Toy Racer

2006-04-05T16:15:00.000-07:00

A camera strapped to a tiny toy track race car.

An omni-cam or wide-angle lens would certainly help.

New Job

2006-03-22T07:43:00.000-08:00

I just left my job at Charles River Analytics, where I was working on automated nighttime outdoor surveillance. I'll be going to iRobot, to work as a software engineer for PackBots.

I am VERY excited by the opportunity, and will relay as much interesting and non-confidential material as I can, erring on the side of caution.

NYC gets more eyes

2006-03-22T07:38:00.000-08:00

From Drudge, an interesting story about NYC beefing up their surveillance system.

Most people might be concerned by privacy issues, and the solution is automation. Humans are very, very bad at paying attention to a single video stream for more than 5 minutes. Paying attention to hundreds would require the manpower of a small army, which the local police just don't have dedicated to the task. So you needn’t worry about someone watching – though the police should be if they expect such a system to prevent crime. It will work by in afterthought or in deterrence.

A computer aided surveillance system, where people and cars are tracking and logged, with learning of truly suspicious behavior, would be significantly better. Also, computers aren't interested in peering into private residences. To stress, these comments are about _public_ surveillance.

IntelliVid is a Cambridge MA based company where I almost took a job. They are working on aided surveillance for "loss prevention" -- retail theft at your local Macy's or what have you. I've seen their system, and it is quite good.

Many companies are working on outdoor automated surveillance, mainly because people like the military, border patrol, police, etc. would all be interested. The best system I've seen is by ObjectVideo, characterized by very few false alarms. Sarnoff is also extremely good.

CROWS

2006-03-10T10:18:00.000-08:00

An interesting description of CROWS (common remotely operated weapon stations), which is a tele-operated weapons system, typically mounted on hummers.

These systems should be entirely remote, deloyable via micro UAVs, much cheaper, and much more common.

Look here for some videos.

“It’s surreal to be in combat and then go buy a Slurpee”

2006-02-24T11:23:00.000-08:00

On the Predator, and its maker, General Atomics Aeronautical Systems (GAAS).

Video of Pleo

2006-02-09T08:51:00.000-08:00

Very cute robot. I'd like to see more tools and more explicit life-forms.

Middle ground with hard-skinned Aibos or effectively useless Asimos is a waste of time.

As far as human interfaces, a cute pet makes much more sense than a bad CGI face on a monitor.

Grand Challenge 2006?

2005-12-27T09:00:00.000-08:00

You can help decide.

"The androids, automatons and artificial animals website"

2005-12-16T12:40:00.000-08:00

Interesting. Check out this video.

Joe Engelberger at IREX

2005-12-03T20:19:00.000-08:00

Behind those flirting eyes and underneath that weird stuff that might be skin, is the future. Or so they think in robot land. Who do they think they are fooling?

Certainly not Joe Engelberger, the father of modern robotics, who emerged exasperated from a visit to the international robot show in Tokyo to stick a prong right into the circuitry of the Japanese robot industry.

"These are toys that are being made," he said during a 'oh heaven help me' plea to an audience of scientists and industry representatives in Tokyo this week.

There are indeed many cure useless humanoids out there, but the research isn't without reward.

Generally, I'd like to see more people try to solve part of the problem first, and then expand the frontier. He gives the plea to develop aides for the elderly. I think a "smart" walker could help.

There probably isn't enough of an engineered environment, which is why most people think that robotic aides are a ways away. Start from something simple: are there stairs? What are you going to do about it?

Get a bit more complicated: how do you handle delicate items like dishes?

I think there needs to be a flexible platform developed that can start by using tele-operation, and move on to more complicated tasks autonomously. It probably shouldn’t be humanoid. It needs cameras and sensors for local obstacle avoidance. It needs fairly dexterous manipulators on arms which can reach a high shelf.

At a presentation about 2 years ago at CMU, I saw Engelberger speak about this same topic. It surprised me that he was asking for funding. Why should just a man need to look hard for what he considers to be an urgent and lucrative application of the field he pioneered?

Big Dog

2005-12-02T07:27:00.000-08:00

In a previous post I just mentioned Big Dog, a 4-legged robotic mule. Last time I tried to download the video, it was cutoff without my realizing it. Now I've seen the entire thing, and I encourage you to check it out!

I'm told there is a stereo sensor head. You would need to do some excellent obstacle avoidance. The human control scheme is the most interesting aspect to me (mainly because I don't know much about legs). Do you use a joystick & line-of-sight? How about tele-operation? Would you need to stabilize the video feed?

I've worked on perception for a similar platform, RHEX, so I'd be interested in seeing what they're planning.

HAL: Hybrid Assistive Limb

2005-12-02T07:17:00.000-08:00

Assitive Suits are very interesting. The system mainly interacts with you, not the environment. Imagine the difficulty: it needs to support 200kgs, but detect when you want to move it. If done well, the tiny force your hand exerts on the suit will make it respond.

Suits will initially be for the disabled and elderly. Shortly, it will be for the military. A bit after than, it'll be for consumer use.

For the military, the main concern is the massive amount of weight soldiers carry. A good quote, from a general: "soldiers today carry 80-140lbs of extremely lightweight equipment".

If they could handle a 500kg load, they could carry water and food for days, in addition to heavier armor, much more weaponry and ammunition. This would make each soldier much more powerful. An alternative is a mule robot like Big Dog.

A huge issue is power. My guess is that in the short term, at least for outdoor operation, they need to be big enough to hold an internal combustion engine. Latter, fuel cells should handle it. If I were to guess, I would estimate the HAL-5 on the right gets 15-30 minutes battery life.

By the way, there is the International Robot Expo (IREX) in Japan on right now. Look for the latest and greatest in robotics.

Sub-Optimus

2005-11-29T06:51:00.000-08:00

"Bad Optimus - A page dedicated to the worst Optimus Prime outfits on the net"

Actually, some are quite good. Is there currently a good robot cartoon to encourage kids to either join the military or go into robotics? I haven't watched any toons in a while.

Defense Tech: "Future Combat": Cuts, or More Cash?

2005-11-18T10:44:00.000-08:00

“They are looking to slip it to the right or kill it,” said a source familiar with FCS options advanced by the Pentagon's office of program analysis and evaluation.

Army officials are working to convince Pentagon leaders, including England, to reconsider cutting or even terminating FCS, the service's only major new-start development program.

Army Chief of Staff Gen. Peter Schoomaker is scheduled to meet Friday with England and again make the case for the program, a briefing that is expected to discuss FCS' relevance to today's challenges.

This is bad for robotics. The benefit of an automated security force is huge. While budgeting issues are serious, the biggest push for robotics shouldn't be cut so quickly.

CBS news report on Robots in Japan.

2005-11-14T06:05:00.000-08:00

In Soviet Russia, the Robots Remote Control YOU!

2005-10-26T08:49:00.000-07:00

Note that my tag-line is not intended to be automating human choices, but mechanizing human thought :)

I should change it.

Seminar on Dangerous Ideas - Harry Potter's next wand

2005-10-26T06:56:00.000-07:00

Sounds like a good talk. They should work on getting lighting to shoot out of the wand.

-----------------------

Harry Potter
Dangerous Ideas Seminar Series Fall 2005
Speaker: Larry Rudolph
Speaker Affiliation: MIT CSAIL
Host: Metin Sezgin
Host Affiliation: MIT CSAIL

Date: 10-25-2005
Time: 1:00 PM - 2:00 AM
Refreshments: 12:45 PM
Location: D449 Kiva

Title: Harry Potter's Next Wand: Should it be digital?

Harry Potter and other wizards preferred input-output device is a wand rather than a laptop, PDA, or cell phone.
This informal presentation will discuss the merits of a portable computer that looks more like a wand or walking stick rather than the current pack of playing cards.

A stick or cylindrical tube can be held in the hand and carried around.
It allows many arrays of sensors to be placed along the length at
precise, equally spaced locations. It seems ideally suited for many
input/output functions.

Many people have strong opinions about form factors, and this proposal oppears to challenge them. Must we continue to make interfaces smaller and smaller?
Must we continue to make them rectangular? Must they all be approved for use on an airplane? Come to the seminar and express your opinion.

Mr Soccer Robot Football

2005-10-25T10:44:00.000-07:00

Robot Soccer!

Technically, this is Remote Controlled robot soccer, as the robots aren't autonomous, but this is an excellent idea. They should make a second gen version with a hybrid of autonomous and RC players.

I wonder how well this will do. If it seems like it would be easy to hack, I might buy one. I wouldn't mind setting it up so that the robots shuffle armies around in a game of Risk.

Velociraptor Dino Not So Vicious

2005-10-21T06:56:00.000-07:00

The simulated velociraptor leg:

The test on a pig carcass:

We need more robots-simulating-dinosaurs.

The Grand Challenge Is About Off-Road Driving

2005-10-12T08:16:00.000-07:00

This is probably a pretty innocent mistake, but many commentators have seen the Grand Challenge only through the lens of automated civilian driving – the kind most people do every day.

Let me just point out that there is a world of difference between off-road dessert driving and freeway driving. There are also huge differences between freeway and city driving.

Dessert driving is very hard, even for humans. Driving on a rocky, dirt road that you’ve never seen before, for those used to nice, paved roads, is a challenge. In most cars, you couldn’t top 20mph, making the average speed of Stanley, 19.1 mph, seem pretty good.

Off-road, there are few mobile obstacles, many large obstacles, many obscured obstacles, very tight turns, steep inclines, no road signs, no lanes (or lane markers).

On paved roads, you have signs, smooth curves, lane markers, barriers on the road edge, reflectors in the lanes and on the side of the road. Unfortunately, you also have other drivers, and sometimes pedestrians. You are also traveling, on average, at much higher speeds. This changes something called “look-ahead distance”, i.e. how far ahead you need to look for threats to avoid them. This is actually the governing reason why high-speed roads are so smooth; even if you could, going around a corner at high speeds means you probably can’t stop by the time you see there is a reason to stop.

Despite these obvious differences, there is a much better reason to differentiate between road driving and off-road driving. The former is much further along! It has deserved comments by the main stream, but didn’t have a “race” to catalyze the media.

For example, look towards CMU’s Robotics Institute’s NavLab project. It is a decades old project. They have cars that can go freeway speeds in suitable conditions. If you really want a revolution in civilian driving, go to their page, watch some movies, and call your favorite car company, asking them to incorporate the technology.

Automated driving is also making big splashes in trucking. This should be expected, as the sensors required are more expensive, but the benefit is higher. Currently they are in the form of safety systems that monitor the driver. They can detect lane drifting, and even weariness in the driver’s eyes. This is a good example and this is a good resource.

"Splashpower wants to cut the charger cord"

2005-10-10T17:26:00.000-07:00

You can use them while jogging, take them into the woods or ride with them on the open sea, but portable gadgets still need to head back to base every time their batteries need recharging. The power cord remains the final connection to the wired world for many devices now that technologies like Bluetooth are replacing data cables — but it too might be going away if a U.K. start-up gets its way.

Splashpower Ltd., established as the result of a business competition at Cambridge University, has developed a wireless charging system that uses electromagnetic induction to accomplish wireless charging of devices.

“It’s basically the concept of creating a magnetic field that goes parallel to the surface of the pad rather than out of the pad and this has many benefits,” said Lily Cheng, chief executive officer and cofounder of the company, speaking at a news conference. “It enables us to deliver a very uniform output across the pad and enables us to make a receiver coil that is very thin.”

This could be huge. A big reason robots lack deployment on par with their abilities is that their abilities are very shortlived due to poor battery technology. In certain environments, most notably the home, this could change if charging were constant.

DARPA Grand Challenge 2005 - Results

2005-10-08T15:49:00.000-07:00

It looks like Stanford's Stanley and both CMU entries have made it past the finish line in this year's Grand Challenge. I'm pretty sure Stanford will get the $2M, but both schools should expect big opportunities from this.

Follow the link for the latest info. Check by here later for a more in-depth analysis of what this means for the military & civilian applications begging for automated transportation.

More robotic education

2005-10-07T13:30:00.000-07:00

Robotic Education is an excellent idea. There are a number of very good things about it.

The interaction could be unsupervised by a human adult; "unsupervised playtime" is supposed to be an essential component to boosting creativity and critical thinking. I'm not sure whether a somewhat un-intelligent robot counts as supervision.

The robot could get an excellent idea of the state of development and potential for growth of the child. Specifically tailored education could result.

Constant availability to answer questions and instruct is a huge bonus. Rather than dumping kids in front of televisions, they will interact with an educator, and like it.

A stronger interest in engineering and robots will appear.

Last but not least: dance lessons.

"Robotic-vacuum maker, BU team up on antisniper device"

2005-10-06T12:05:00.000-07:00

IRobot demonstrated the system, called REDOWL (for Robot Enhanced Detection Outpost with Lasers), at the Association of the United States Army convention in Washington yesterday. Testers struck pieces of metal to simulate gunshots. REDOWL quickly aimed its infrared camera and laser rangefinder at the source of the noise, just as it did in tests at a Medfield gun range.

REDOWL is based on iRobot's PackBot, a battery-powered lightweight robot already in active service with the armed forces. PackBots are used to explore dangerous terrain or enter buildings to search for booby traps.

Glenn Thoren, deputy director of the Boston University Photonics Center, wondered if an antisniper system could be mounted atop a PackBot. He used laser rangefinder gear from Insight Technology Inc. in Londonderry, N.H., and sound-detection equipment developed by BioMimetic Systems, an acoustics start-up company founded by Boston University. The REDOWL also includes a Sony digital camera that can zoom in on distant objects or people, and display infrared images at night. When REDOWL's microphones detect a gunshot, the device calculates the source of the sound, swivels the camera, illuminates the target with either visible or infrared light, and uses a laser to calculate the range.

Early Non-Invasive Brain-Computer Interface

2005-09-30T07:01:00.000-07:00

Computer scientists have created a hat that can read your thoughts. It allows you to stroll down a virtual street. All you have to do is think about walking.

Called a brain-computer interface, the device detects activity in certain brain areas linked to movement, and uses the signals to mimic that movement in a virtual world. The technology could one day help paralysed patients to move robotic arms, or help sufferers of motor neuron disease to type out words on a virtual keyboard.

"Clever artificial hand developed"

2005-09-09T08:12:00.000-07:00

As engadget pointed out, it had to be holding a can of Healthy Beans, didn't it?

I wouldn't mind if it were gripping one of those stress balls, especially if it were a sphere painted to look like the earth. Now that would send a message.

The LUCAS CPR bot

2005-09-07T13:19:00.000-07:00

Katrina will probably make people desire more self-sufficiency. A bot like this that does CPR would help. I would imagine adding a touch screen and mobility would allow for a home health care bot, first used for the elderly. It could remind you to take your pills, contain an expert system to ask about possible drug combinations (e.g. "can I take aspirin with my heart medication?"), and save your life by performing CPR or defibrillation.

Generally, I'm surprised the first step hasn't been achieved, e.g. MS Doctor, a cheap expert medical system with a simple interface. Add froogle and a recommendation system, and I would have liked to ask it questions like "Do omega-3 fatty acids work, and what is the best place to get them?", or "Describe the most recommended LASIC procedures in my area".

Modulobe: Virtual Creature Simulator

2005-09-06T06:25:00.000-07:00

Awesome collection.

I know of projects at CMU which simulate humanoids in a realistic, dynamics enabled simulator.

So why can't they make a humanoid with appropriately trained gaits on compliant legs?

Grand Challenge II

2005-09-06T06:23:00.000-07:00

Someone is going to win this year. This list is fairly shallow, but it's important to see what the majority see.

For a better list of teams, and more about the race go here.

Koizumi vs. Robot

2005-09-01T11:16:00.000-07:00

Click the top left blue link.

MS Student

2005-08-29T11:05:00.000-07:00

Microsoft isn't the only one making educational software. I very much look forward to this kind of software becoming more common as the performance improves.

Simple programs to animate examples and spit out well rehearsed explanations will be trivial.

More advanced programs that can dynamically assess which parts of a problem are not well understood will help.

Situationally aware programs which have an excellent idea of what the student is reading and maybe thinking, in addition to long term performance tracking, will be amazing. At the very least, a program that can discipline a student for not paying attention will be a roll reversal.

Also, more specific domain knowledge will allow an education frame work to be easily applied to a new area. I am very hopeful about the potential benefits. Bringing world-class education to anyone who has access to a computer (which, considering public libraries, is everyone in America) will be revolutionary.

wakamaru

2005-08-29T10:54:00.000-07:00

Living with the family
"wakamaru" lives in accordance with his own daily schedules as well as those of its owners, and it can update these schedules based on contact with the owner.
Speaking with the family
Not only does "wakamaru" respond to actions from people like other conventional robots, but it also takes the initiative in speaking to the family based on the information he obtains from contact with the family.
Its own role in the family
"wakamaru" connects itself to the network to provide necessary information for daily life. It looks after the house while the family is absent, watches out for unusual conditions, and is convenient for the life of family members.

Digital People

2005-08-26T12:39:00.000-07:00

Interesting Slashdot review of "Digital People: From Bionic Humans to Androids".

Robo Maxx Floor Cleaning Robot

2005-08-26T06:41:00.000-07:00

Engadget links to yet another vacuum bot. When will they make bots to clean my tub?

I just bought a Roomba Discovery. I like it a lot, but it isn't very intelligent. It is plenty for the price though.

Fly pentop computer

2005-08-24T13:58:00.000-07:00

How times have changed. Back when I was growing up in the late 70s and 80s, lots of parents thought Scrabble was all the stimulation children needed to excel in school. Nowadays, adults worry that their kids won't do well unless they own a laptop, a cell phone, and a personal digital assistant (PDA). And soon, perhaps, the Fly pentop computer will be added to that list.

Developed by LeapFrog Enterprises (LF ), a maker of educational toys, the gadget is just what it sounds like: a talking computer hidden within a pen the size of an electric toothbrush. Due to become available at stores like Wal-Mart (WMT ) and Target (TGT ) for $99 in mid-October, the device is aimed at 9- to 14-year-olds. What can they do with the Fly pentop computer? Kids can use it as a calculator, keep a calendar, create and record music, and play complex logic and geography games -- all features I tried and found fun and educational.

This sounds like an excellent educational tool. I'm looking forward to more computer based education.

New Asimo

2005-08-24T06:21:00.000-07:00

Why would Honda want to spread an image of Asimo trying to run down an old man?

The locmotion looks good. Too bad it isn't bio-mechanically intelligent.

Enhanced hearing

2005-08-18T12:24:00.000-07:00

"This article in Wired, the U.S. technology and futurism magazine, says manufacturers are now producing aids that even someone with "perfect" hearing can use to enhance the experience of certain sounds. An example of how a medical technology ended up becoming almost a luxury product like colored contact lenses."

Yes, I am hoping this is a trend. Technologies used to free people of their disabilities will be adapted to enhance those without them. This will probably make an "us" and "them" view of a world with advanced robots a bit limited.

Elvis the robocat - Engadget

2005-08-16T07:00:00.000-07:00

"You just gotta see it to believe it. Carlo Bertocchini’s cat, Elvis, was involved in a truck accident that left him without use of his rear legs. Carlo built little Elvis a moving platform to regain some of his lost mobility, which the cat actually pilots by depressing two control buttons. We’re impressed that cats be trained to do such things but more importantly, we can’t stop laughing over this video."

Watch the video. It doesn't really look like the cat knows what it's doing. Also, if only the back legs were hurt, why is the entire cat in a box?

Wouldn't it just make sense to have his back legs in a wheel chair of sorts? You wouldn't even need motors. I guess the point is that we couldn't laugh at a cat in a big yellow box.

John Carmack's QuakeCon 2005 keynote

2005-08-16T06:03:00.000-07:00

"Carmack said he considers the prospects for the upcoming physics acceleration chip on the PC iffy, because physics presents a very fundamental problem that graphics doesn't have: it isn't easily scalable for level of detail."

I wonder if the primitives can be used easily for dynamics modeling. This would be interesting both from a development perspective, as a simulator, and from a deployment perspective, where the models could be used to make more accurate predictions (given a good model of the world) about a robots path, for example.

Thin skin will help robots 'feel'

2005-08-16T05:56:00.000-07:00

Japanese researchers have developed a flexible artificial skin that could give robots a humanlike sense of touch.

The team manufactured a type of "skin" capable of sensing pressure and another capable of sensing temperature.

These are supple enough to wrap around robot fingers and relatively cheap to make, the researchers have claimed.

The University of Tokyo team describe their work in the latest issue of the journal Proceedings of the National Academy of Sciences.

I'd like to see dense tactile sensors used in combination with camera based object recognition. That seems like the way a toddler does it. I'm sure they would be quite complimentary, one gives you texture and the other gives shape and color, roughly.

Perhaps an example from the wonderful world of fruit would help.

Nectarine:

Peach:

The under-reported good news about productivity

2005-08-15T06:01:00.000-07:00

This might not seem related to robotics, but a common concern is that robots will "replace us". In the short term, this is to mean that they will "take our jobs". The meme is based on the idea that if a job takes fewer people to perform, those unneeded are left unemployed.

This is not just false in theory, but wrong empirically as well.

As there is no limit to human desire, there is always a push to do more. Gains in productivity should free resources such that those whose skills would be wasted making widget X are better-used providing service Y or widget X.2.

In addition:

There will always be some pessimists who think that productivity growth is a bad thing, reasoning that if one person can do the work of two, the unnecessary second worker will become unemployed. The record of history on that hypothesis is extremely clear, however. U.S. workers today produce more than three times as much per hour compared with their counterparts 50 years ago, and even so, the unemployment rate today is the same as it was in 1950. Instead of putting people out of work, what productivity growth has always meant in practice is a rising standard of living for everyone. No other statistic may be as important for determining long-run economic welfare as productivity.

This makes sense. People can produce more (or machines can do more for people), in a given amount of time, and more things get made.

___

By the way, in the long term, I think "embrace us" or even "embrace them" is a more appropriate description than "replace us" for the 30-50 years from now when human-scale artificial intelligence is developed.

Scribbler

2005-08-13T20:06:00.000-07:00

"The Scribbler is a fully programmable, intelligent robot with multiple sensor systems that let it interact with people and objects. It navigates on its own as it explores its surroundings, and then reports back about what it senses using light and sound."

Retails $100.

University of Minnesota Center for Distributed Robotics

2005-08-12T06:15:00.000-07:00

A good range of smallish bots. Some nice videos too.

P.E.A.R.T. - The Robotic Drum Machine

2005-08-11T06:14:00.000-07:00

Awesome. I suppose a synth would do, though.

Artificial actors

2005-08-04T05:55:00.000-07:00

I have some comments over at Marshall Brain's Robotic Nation Evidence on a post about Robotic actors.

____

Part of the point is that many hundreds of people do a lot of work into software that makes life easier for people, even just one person.

I used to love the advance of video games where one game's "cut-scene" was the next generation’s game's live rendered graphics. A perfect example is Tekken 3. The cut-scenes were the typically smooth CGI with worse in-game graphics. The platform was the PlayStation. To show the power of the PlayStation II, they showed what looked like the beginning of a cut scene, but the characters were controllable and could be moved around the room from the old cut-scene.

Recently, with Doom 3 providing real-time bump mapping (an amazing achievement), I can only expect this trend to continue. The only reason it wouldn't is that the graphics are getting so good (and so much work is required to make the engines) that game makers are creating cut-scenes which are just scripted real-time rendered graphics, so this leap-frog effect is lost. Soul Caliber II comes to mind.

Then again, an excellent possibility of this is to have hybrid game/movies, where cut-scenes are merged seamlessly with game-play. This is getting more common, and allows games to have the delight of immersion and user control, in addition to the passive enjoyment of watching a directed and well-produced short film. I’m pretty sure “Enter the Matrix” was like this, but I never played it.

Marshall's post about HalfLife II’s progress is very relevant here.

Brain cells on demand

2005-06-29T14:18:00.000-07:00

Regenerative medicine scientists at the University of Florida's McKnight Brain Institute have created a system in rodent models that for the first time duplicates neurogenesis -- the process of generating new brain cells -- in a dish.

Writing in today's (June 13) Proceedings of the National Academy of Sciences, researchers describe a cell culture method that holds the promise of producing a limitless supply of a person's own brain cells to potentially heal disorders such as Parkinson's disease or epilepsy.

I love the was diseases are the first thing mentioned for such applications. I personally find the potential affects on society once intelligence can be increased without consequences to be more important. Unfortunately, like anti-aging, there is a certain societal aversion to playing god like this. FuturePundit often comments on the potentially huge positive effect an increase in average intelligence of even a minority of a population can have on GDP.

It would also be interesting to see what affect this has on growing wetware interfaces to computers.

Roshambot

2005-06-28T06:41:00.000-07:00

AI dawns.

Extracting Video from the Brain

2005-06-23T06:44:00.000-07:00

Awesome.

Stanley claims the decoding algorithm is simple since each point in space can be reconstructed at high resolutions from 6 to 8 pairs of cells. The more neurons are tracked, the higher the quality of the reconstruction. The following still frames from a video (top) are actual reconstructions (bottom) from the activity of 177 cells in a cat’s brain.

Check out his research page.

Grand Theft Auto Meets Robocop

2005-06-20T15:15:00.000-07:00

An excellent vision application.

he Plate Hunter is based on optical character-recognition technology originally developed in Italy for sorting letters and parcels.

"Five to six years ago (Elsag) had really perfected postal sorting and realized that if they could read a postcard handwritten at 90 miles per hour, they could read license plates," said Windover.

A similar system has been used by the Italian carabinieri for the last three years, and is mounted on 3,000 Italian law enforcement vehicles.

Remington-Elsag has also developed a fixed system. Mounted on the side of the road, the reader can scan vehicles moving at up to 75 mph, with recognition rate in tests exceeding 95 percent, the company claimed.

Windover said false positives are "virtually nonexistent."

"In California, where we have the most experience, false positives are rare, occurring less than 1 in 100,000 reads," he said. "Importantly, 100 percent of all alarms from any read are verified by the operating law enforcement officer prior to a traffic stop or other action."

And the money shot: "It doesn't require the driver to do anything."

Japan Dreams of Robot Moon Base in 2025

2005-06-20T10:30:00.000-07:00

Most people don't know it, but Russia sent robots to the moon in a much more efficient but much less inspiring push to the moon compared to Apollo.

Robots just make sense. You can send twice as many when you don't need human maintenance equipment, and because the moon is closer than Mars, they can be tele-operated more efficiently.

Ultra-Lifelike Robot Debuts in Japan

2005-06-15T06:43:00.000-07:00

Brain cells grown in Lab

2005-06-14T04:59:00.000-07:00

This is huge. First it could cure many degenerative brain diseases. Then it can be used for enhancement. Folks have been trying to do this for some time, but it seemed that growing new adult brain cells was impossible.

"The Man with the Bionic Ear"

2005-06-13T10:40:00.000-07:00

very interesting

Excellent Tech

2005-06-09T06:07:00.000-07:00

"Grocery wholesaler James Hall has cut costs and administration time by investing a six-figure sum in a voice technology system to help staff pick out orders."

This is also discussed here, where the blogger thinks this will lead to a rise of robotics, in a very negative way.

Both CMU entries reach robotic race

2005-06-08T05:27:00.000-07:00

I'm looking forward to the competition. Someone will win this year. Next year there will be a huge push for mass production. The year after that, if there is a war going on, this technology will be used to automate everything from convoys to ambulances.

Unfortunately, with the regulatory impulse in America, it will be some time before transportation (usually the most expensive component of commodities like food) will benefit.

Call to bullpen: 'Give me R2-D2'

2005-06-08T05:22:00.000-07:00

While Pirates pitcher Oliver Perez's job is in no danger, coaches and others who pitch batting practice before games could one day be relegated to the bench by a robot.

A preview could come tonight, when the S-3 Platform Robot, built by Aliquippa-based RoPro Design and Beaver County Area Vocational Technical School, will deliver the first pitch as the Pirates square off against the Baltimore Orioles on "Robotics Night" at PNC Park.

So automated pitching machines don't count? I guess they never use them in a game. They should have a 0th inning, where 3 batters from both teams compete in hitting against a really powerful pitching robot. This could be used both to break ties and humiliate merely human players.

Scooba makes mopping obsolete

2005-05-31T06:15:00.000-07:00

I would worry about this new iRobot wet-vac. It has a far smaller market size, and seems to have higher maintenance.

One big problem is that getting stuck might mean spilling a great deal of potentially staining cleaning fluid in one spot.

Also, one task like vacuuming can be done at random because there is no series of events. Just vacuum. This guy vacuums then wet-cleans then mops up. What happens to the dirty water receptacle when a missed wet-hair-ball gets trapped in there?

I personally don't mind mopping. It's fast. Scrubbing bathroom tiles and bathtubs are what I really hate doing.

I do plan on getting a regular Roomba sometime soon.

FEndorphin : 3D Character Animation AI

2005-05-31T06:08:00.000-07:00

Add dynamic models to biomechanical 3D models, mix in some genetric algorithms, and output realistic movement.

Shadow Robot Company:Commercial Hand

2005-05-24T07:19:00.000-07:00

This seems very, very imporessive. While undoubtedly tele-operated, this dexterous hand looks like it has a great deal of potential. Check out the video here.

[via FutureFeeder]

Mech

2005-05-09T07:05:00.000-07:00

I'm not sure if this is real. I certainly would need a video, but it won't play for me.

Update: try here, with a bad video here. Obviously not done yet.

Cyc

2005-05-09T07:00:00.000-07:00

Cyc is about to go online. THe company that has sometimes been called the biggest AI pork spending project ever will face the rath of thousands of people trying to trick this chat-bot .

Supposedly it isn't based on simple rules like other brain-dead chat bots like ALICE and Eliza.

Adidas 1 - a smart shoe

2005-03-18T07:40:00.000-08:00

At $250, it better be. I'm no expert in wearables, but I'd be interested in any device that could facilitate passive communication in a ubiquitous computing environment.

Human identification is so hard, I'm certain some device will eventually make it secure and easy before the distance retina scanning or face recognition is adequately robust. Once the conversation starts, logging into your computer, paying for things, and unlocking your front door should all get a bit easier.

In addition, the more intelligent devices around you, the more information each device has about its surroundings. This is not only incentive to make everything from your sneakers to microwave "smart", but also avoids the general object recognition problem.

On a side note, why don't foods come with bar codes that a microwave or toaster oven can read and then properly cook the foods? Microwave popcorn is the first to come to mind. This wouldn't even involve something like an RFID, just ink.

Walking Videos

2005-03-18T07:29:00.000-08:00

John Deere has made a 6-legged walking robot the size of a truck. Follow the link for some videos, or go here for more information.

Cameras everywhere, please

2005-02-22T15:18:00.000-08:00

World Tribune.com -- Spy cameras to track spies in Korea's mine-infested DMZ::

I love these stories about the exploding spread of surveillance cameras in all sorts of hairy places. They are cheap, and will be multiplied in value in the short term as computer automatically track events, so a human operator can look at many feeds. Eventually the ratio won't be 3-1 cameras to operators, but 30-1 and then 3000-1, with many decisions made automatically.

With military situations like this, there will be a human in the loop for some time. For less deadly matters, like traffic violations, the humans are already gone and won't be coming back.

People concerned about privacy should pay attention to the facts that this is all in the public domain and that people (not computers) are malicious.

New Humanoid

2005-01-06T07:31:00.000-08:00

Under the misleading title 'South Korea claims world's smartest robot', this article describes a new humanoid 'network-based human-like robot'. I'm not sure what is human-like in having a brain communicate wirelessly with a body, but learning is a good thing to stress in developing humanoids, as is the easily scalable computational power of a backroom server.

I generally enjoy development of humanoids, because they can operate in our world more easily, but I do not like the legs, which are totally the wrong approach. There is no 'bounce' in the step that allows for a great deal of energy conservation and true running.

What is entirely missing is a large humanoid development platform for researchers. Robots like ASIMO prove their worth only in corporate branding, because you can't really control the robot as a robotics researcher would want. I wish these companies would learn something from open source software: letting a talented user peak under the hood will only make your product better.