Kiva Robots Helps Company

Zappos.com is going robotic with the use of robots to handle some of their warehouse work. The company has bought a Kiva Mobile Fulfillment System from Kiva Systems. The system comprises of several orange-colored robots, called Kiva workers that can handle a plethora of different jobs involving inventory in warehouses.

Zappos will use these robots to speed up the process of finding, packaging and shipping a customer’s order. The robots will work in the company’s Kentucky warehouse and will streamline the process of getting customers’ shoes (which is Zappos’ main product) shipped.

Once an order is placed, the robots get a signal from a Kiva server over a Wi-Fi network, which tells them where to go and what box (or pod as Kiva calls them) to pick up. The robots can stack several pods if needed and once their list is filled they bring the pods to a stationed human worker. The worker then takes what he/she needs from the pod, assembles it and places it in a box. That box is then placed on another Kiva robot, which brings the box to the correct location to have it shipped out to the customer. Meanwhile, another Kiva robot is already waiting for the worker to pack and give it the next order.

With several of these robots the company can cut a lot of employee costs and also make their operation more effective. The software which drives the Kiva robots is really powerful and allows companies to use the robots without much hassle once everything is set up. Look for Kiva to grow in popularity as their product gets more exposure in the market.

Discounted dino robot Pleo still a waste of money

UGOBE has been celebrating the 1-year anniversary of their PleoWorld website which has become a central place for Pleo owners to hung-out and talk about the many and better alternative ways they could have spent their money if they had not wasted it on a Pleo. Yes, it is no secret that I may be one of the few (if not the only one) people who has said since even before the robot Pleo made its market debut that it was going to be a massive failure destined to disappoint everyone who buys it.

At any rate, UGOBE will be selling Pleo at a discount for a few days as a birthday promotion for their website. You can get a $100 discount if you use the promotional code “PLEOWORLDISONE” when you buy one of the robots. Unfortunately for everyone, selling the robot at a lower price still doesn't make it into a life-form or even add any of the overhyped intelligence (with emotions and online learning nonetheless) promised to consumers. If you have $235 to spend, then I suggest that a better use for your money would be a trip to your local theater to watch Pixar's Wall-E a few times and why not buy a few copies of the DVD when it is released in a few months. I guarantee that you will receive many more hours of entertainment from Wall-E than you ever could from owning a Pleo!

Robots join the war against woody weed infestations

A new project was recently funded at the University of Sydney with the goal to develop unmanned aerial vehicles (UAVs) for the efficient detection and eradication of woody weeds. Professor Salah Sukkarieh expert in UAVs and currently a member of the Australian Centre for Field Robotics (ACFR) will lead the project with a focus on the development of data fusion and recognition algorithms for the detection of woody weed infestations and their eradication via the pinpoint delivery of herbicides. The team will use a pair of robotic UAVs, one fixed wing and one hovering. The former will fly reconnaissance missions to detect infestations and the latter will deliver the herbicide in a targeted manner. ACFR already has two such vehicles in use and what remains to be done is the development of the algorithms as mentioned earlier.

This is an interesting application for UAVs considering that such robots are predominantly used for military purposes either for surveillance or the delivery of explosives on enemy positions. Currently, weed infestations are a major problem in Australia's open grasslands amplified further by labor shortages. The Australian government is dedicated to funding programs that can help reduce the country's woody weed problem via their Defeating the Weed Menace program which also includes work on combating aquatic weeds.

WowWee RoboRemote To Control All Robots

WowWee has released the RoboRemote, which can control all robots made by the company, including those that are still to be released. Yes, it’s not actually a universal robot remote, since it only works with their products, but hey, that’s still good if you have more than one WowWee machine.

The remote can be hooked up to any PC using XP or Vista (no Mac yet) to be programmed. The computer hookup is also used to install updates. Basic commands are already preprogrammed, so the remote can be used out of the box with most robots, although for more advanced commands you may need to reprogram some keys. The software that comes along with the remote is very powerful and easy to use, allowing users to not only program commands, but also issue time delays, and even a chain of commands. These tools allow users to create complex programs for their robots, which can be triggered by one simple stroke of a button.

The remote has four programmable buttons, which can do different things as well as a shift button. There are three different shifts available (all programmable), which can be used for different commands. An LED in the center of the remote changes colors based on what shift you are on as well as whether or not a button issuing a command is being pressed.

Initial critic reviews have said that the product is very easy to use. At only $20, it is a great utility for anyone who has more than one WowWee robot.

A robot that plays air-hockey… and wins

Nuvation, an electronic engineering company, has created an air-hockey playing robot to display the skill and brilliance of the company’s engineers. The machine is essentially an industrial robot rigged with a motion sensor through which it tracks the puck (which has a silver decal on it that acts as a tracking device.) The machine is programmed to stop the puck from entering its own goal and after stopping it to, of course, shoot it in the other direction at its human opponent’s goal.

The company claims that the robot is able to win 90% of its games and at a recent display that statement seemed pretty accurate. The robot was able to defeat most of its opponents, often with stunning shut-outs. Since the robot has a motion sensor and the technology to detect where the puck will go, it often makes crazy saves that ordinary players wouldn’t be able to make. For example, humans often move the paddle so that it is centered when it hits the puck, but Nuvation’s creation can detect how the puck will bounce off ahead of time so it often doesn’t even have to center the hit.

The calculations and adjustments the robot makes are fascinating and only someone that really knows what they are doing would be able to defeat the machine. The engineers probably know the weaknesses, but they wouldn’t reveal them when asked. And with that we can all add one more thing that robots are better than humans at.

Rodney Brooks: I am a robot and so are you!

MIT Professor and celebrated robotics researcher and pioneer Rodney Brooks recently wrote an article for IEEE's Spectrum magazine presenting his view on the Singularity. The Singularity is the idea that there is a well defined and fast approaching moment in time when artificial intelligence will match and eventually start to surpass human intelligence.

Brooks starts the article by expressing his view that humans are essentially robots made of organic matter and as such it is very much conceivable that machines that are as capable and intelligent can be constructed from non-organic matter; in other words, there is nothing special about humans.

Brooks gives his view of how the Singularity will come about not as an event that will occur at a single point in time but rather as a gradual process over a long period of time. Brooks explains,

I don't think there is going to be one single sudden technological “big bang” that springs an Artificial General Intelligence into “life.” Starting with the mildly intelligent systems we have today, machines will become gradually more intelligent, generation by generation. The singularity will be a period, not an event.

He talks about how this transitional period from human to artificial intelligence will come about as researchers start to build AI systems with the intelligence of small kids (for example, the object recognition abilities of a 2-year old child and the language understanding skills of a 4-year old child) and progressing forward until a computer with the intelligence of an adult human being becomes a reality. He also emphasizes his conviction that true AI can only come about when algorithms are coupled with a body capable of experiencing the physical world. I have no argument against that except maybe my belief that we might be able to build machines that sense the world in ways that humans cannot; this could lead machines to acquire a better understanding of how the universe works something that we, humans, can only do indirectly (if at all) by constructing and utilizing special instruments such as microscopes and telescopes.

These are only a couple of the main points that Brooks addresses in this well thought-out article; another issue that his spent considerable time discussing is whether the human brain can be simulated by a digital computer or if a different kind of computational engine is needed in order to achieve human-level intelligence. This is definitely an open question and an argument that I often make to derail those who believe passionately that faster computers inevitable will lead to machines with human-level intelligence.

The entire article is available online and I highly recommend anyone with an interest in robotics give it a read.

The full article: I, Rodney Brooks, Am a Robot.

NVIDIA 9800 GTX+ is now PhysX-enabled

It looks like NVIDIA's acquisition of Ageia, makers of the PhysX physics simulation engine, has finally come to fruition as it was announced recently that the 9800 GTX+ GPUs will have this feature enabled via a driver upgrade. To be more precise, installing the most recent version of the ForceWare driver will enable the physics calculations on the NVIDIA hardware. I should point out that the GTX+ is essentially an overclocked GTX card so you get the same 128-core system running at a faster speed; the older cards can also have the physics engine enabled using the same NVIDIA driver. So, to summarize, the new GTX cards come with all the bells and whistles people expect from a modern GPU including fast (and, in fact, faster than previous models) clock speeds and lots of high-performance memory.

NVIDIA introducing PhysX-enabled GPU hardware is great news for robotics and AI research for several reasons.

First of all, we no longer need to purchase a separate piece of hardware for physics simulation since a single NVIDIA GPU can do both high-performance 3D rendering and physics calculations. Second, these new cards are very affordable costing less than $200; who would have thought 10 years ago that a couple of hundred dollars would buy you so much horsepower! Finally, let us not forget that PhysX is the the simulation engine used in the Microsoft Robotics Developer Studio. In other words, there already exists a good tool for creating realistic environments and testing AI and robotics systems driven under the wood by what is probably the most realistic physics simulation engine in the market today.

It is well known that NVIDIA is slowly moving into Intel's territory (and vice versa) trying to create a new massively parallel and affordable computing platform. Single core systems went out of fashion many years ago. CPU maker Intel (and AMD) have been slowly introducing their own large multi-core systems (in the lab they have silicon with more than the 4 cores we can buy today at the local computer store.) Considering that some very important algorithms can easily be parallelized (for example, heuristic search planning algorithms,) such developments are more than welcomed by AI and robotics researchers. Moreover, there is much research and development put into developing new languages that would make it easy to write software for such computer chips; difficulty of programming is an obstacle that needs to be overcome soon if we are to take full advantage of these new hardware developments.

Australia's robot clarinet player

This is the first I heard of this but apparently there exists an international robotics competition that is all about creating robot musicians. This year, the Artemis Music Orchestra competition took place in Athens and the winning entry was from Australia. But before we talk more about the robot musicians, this is what the competition is all about according to the information provided on the official website.

The Artemis Music Orchestra is designed to encourage understanding about Embedded Systems in a broad audience and especially young. The ARTEMIS Music Orchestra challenges participants to create devices that play real musical instruments. The contest aimed at higher education schools and universities.

Researchers from NICTA and UNSW developed a robot clarinet player which performed better than a Dutch guitar-player and a Finnish piano-player. It is hard to appreciate how good the robotic clarinet is without actually enjoying one of its musical performances (The Flight of the Bumblebee) so check out the video below (it includes a description of the competition and some insights on what makes building such a music-playing robot hard.)



Finally, below is a video of the second place guitar-playing robot impressing the ladies with an acoustic guitar.

V-STAR: The military's flying Humvee unmanned flying vehicle

If you want to win a war, you have to make sure that the proper supplies, i.e., water, food, and ammunition, can reach your soldiers as needed. For those who have followed the recently conflict in IRAQ might remember how the speed of the USA's advance towards Baghdad was to a large degree constrained by how quickly supplies could be transported to the front line. During the occupation that followed and continues until today, getting supplies to the many bases and soldiers out on missions is a huge nightmare for the military. It is no wonder that events such as the Grand and Urban Challenges were organized in an effort to accelerate the creation of technology that could improve the supply chain in war zones.

Frontline Aerospace has now designed a new unmanned flying vehicle (UAV) named V-Star for the purpose of supplying troops in combat situations. The company refers to the UAV as a flying Humvee because of its size, load capacity, and functionality as a transport for supplies. The autonomous aircraft can take-off and land vertically while it can fly anywhere from 600 to 1000 miles and carry 400 pounds of cargo. The V-Star is about the size of a large SUV (which once again explains the nickname flying Humvee.) Finally, the V-Star can fly as high as 15,000 feet and it makes less noise than a helicopter due to its propeller design.

Frontline Aeorspace envisions the usage of their UAV in a scenario in which a soldier could request supplies via a wearable computer system and have these supplies delivered to his location by the robotic vehicle without endangering the lives of more soldiers.

PS: I should mention that this vehicle is not in production but only a concept.

Image is copyright Frontline Aerospace.

NASA tests out next gen lunar robot

Where does NASA go when it wants to try out the next generation of robotic vehicles being designed to trek across the moon? If you guessed “Washington state” you must have been one of the researchers involved with the experiment!

Up until last Friday seven teams from the National Aeronautics and Space Administration were testing out several designs of lunar robot vehicles near Moses Lake in Washington. "The goal was to gain hands-on experience with specific technical challenges anticipated when humans return to the moon by 2020, begin to explore the lunar surface and set up outposts," Test Director Bill Bluethmann of NASA's Johnson Space Center in Houston said of the two week trials. The next gen of space robots got to test out their traction on sand dunes in the region around the lake as well as during extreme swings in temperature and sand storms. NASA is testing several different kinds of lunar rovers including smaller scout vehicles and larger craft designed to carry supplies, equipment or even cranes to be used by the astronauts. The Constellation Program aims to have American astronauts return to the moon by no later than 2020. Already tests are underway for the multipurpose Orion spacecraft, to be used to transport astronauts to the moon.

Now that the tests in the field have been completed the scientists and engineers will go over their findings in preparation for their next round of site testing scheduled for this coming October.

Pomi: A robotic penguin pet

Robot pets are a big market and everyone knows this. A robotic pet is a product that fits nicely in the space between a robot toy and a real pet. Such robots have the potential to be very entertaining, real companions with therapeutic powers and without all the mess of a real pet. It is no wonder then that many companies are trying to come up with the first killer robot application in this space. Pomi is from South Korea and also the Electronics and Telecommunications Research Institute's (ETRI) latest effort to make a splash in the robot pet market.

As you can see from the photo at the top left cornet, Pomi is a robot penguin with a screen on the chest and lots of plastic all over. Its name stands for Penguin Robot for Multimodal Interaction (Pomi) and it has been chosen in order to better communicate and also emphasize that the robot is meant to interact with kids using multiple modes including speech, video, and apparently smell. I don't know how the smell is going to work but let's hope that the engineers had enough sense to only include pleasant smells in its arsenal. It could be that a future application for Pomi would include missions in war zones during which the robot will be given the task of gassing enemy troops (a scary thought to say the least.)

As usual the researchers did not fail to mention how the robot will emulate human emotions and yada yada yada. I say this because it seems that somehow emotional robots have become the speak of the town even though these so-called emotions are really nothing more than a couple of hardwired responses that some people apparently find cute and get all excited about. Robotics researchers really need to get back to working on solving the hard problems that really matter and can help advance the field making useful robots instead of putting all their efforts on pleasing the journalists with silly talk!

The Roadrunner supercomputer simulates the human visual system

It takes a supercomputer to really simulate one of the most complex parts of the human brain, the visual cortex, and the scientists at Los Alamos have one operating for only one week but doing more than any other computer in operation today. The supercomputer nicknamed Roadrunner can execute 1 million billion instructions per seconds (petaflops) making it the fastest supercomputer in operation today; it more than two times faster than the previously fastest offering by IBM. Roadrunner is built using AMD Opteron dual-core processors each with four PowerXCell 8iTM processors for an aggregate of more than 100,000 cores to the tune of $133 million; the supercomputer like many others in the past was engineered by IBM.

So what are the scientists using this computational workhorse for? Apparently, one of the many programs they had running on it this past week was PetaVision which is a huge simulation of the human visual system.

PetaVision models the human visual system—mimicking more than 1 billion visual neurons and trillions of synapses. Neurons are nerve cells that process information in the brain. Neurons communicate with each other using synaptic connections, analogous to what transistors are in modern computer chips. Synapses store memories and play a vital role in learning.

Synapses set the scale for computations performed by the brain while undertaking such tasks as locomotion, hearing or vision. Because there are about a quadrillion synapses in the human brain, human cognition is a petaflop/s computational problem. (source)

Lots of effort in computer vision research today goes in creating image processing algorithms that try to achieve performance at the human level with as little computation as possible; the reason for this is obvious: Computers are not fast enough. In fact, I am actually working on an object recognition system and all my efforts go on minimizing the number of pixels in the image that I consider for the recognition task. If I process all the pixel data then I'd have time to get another PhD before a single trial is done. With faster computers (okay, so you won't have a desktop machine supercomputer with petaflop performance in your home any time soon,) such issues will essentially become less important. It reminds me of what has happened in the computer graphics field after the popularization and performance increase in GPUs. Eventually, the simple and easy to code algorithms will become the focus or computer vision research given the abundance of CPU cycles while the more mathematically advanced algorithms will be put aside.

In conclusion, I have to say that I am glad that one of the problems the scientists are actually experimenting with on Roadrunner can give us better insight about how the most complex part of the human brain works. I say this because it is probably obvious that most cycles on this machine will go on computing more effective ways to 'dispose' of people in the many current and future wars. It is sad that such great machines are used to both discover new knowledge and at the same time destroy fellow human beings.

Roadrunner photo credit: LeRoy N. Sanchez, Records Management, Media Services and Operations

Robot REEM-B video

Below is a video of Pal Technology's REEM-B humanoid robot that we wrote about earlier. It sounds like a great machine but I still think that Honda's ASIMO has the edge on this and all other humanoids available today. ASIMO had the same speech and vision recognition abilities since years ago including autonomous navigation, walking on a slope, climbing up and down stairs (and some times falling down the stairs,) running, interacting with people using speech, vision, and body gestures, and even directing the Detroit Symphony Orchestra. That's not to say that REEM-B is not a great robot but we should not rush ahead and start calling ASIMO obsolete!

Anyways, enjoy the video.

Reem-B robot unveiled at Dubai press event

Honda’s Asimo robot may have met his match. The world got its first official look today at Reem-B, a human-sized robot capable of performing complex movement, walking and recognition of voice commands. Developed by Dubai-based Pal Technology, Reem-B pushes the envelope further for this kind of mechanical assistant. In attendance at the ceremony was Shaikh Mohammad Bin Zayed Al Nahyan, the Crown Prince of Abu Dhabi and Deputy Supreme Commander of the UAE Armed Forces, who watched Reem-B walk, listen to his voice and execute his commands.

Reem-B has voice recognition software, the ability to discern faces as well as infrared and touch sensors in its hands. It can carry up to 13 kilograms. Most surprisingly Reem-B only weighs 60 kilos because it’s made of aluminum.

Pal Technology now wants to mass produce Reem-B in the near future. “The robot will one day assist us in guiding and help mainly the older people, those who have mobility problems, in hospitals," commented David Faconti, the project leader of Pal Technology at the Dubai press conference. But if you are already thinking about placing an order for your own Reem-B, you better hang on to your checkbook: the final price and delivery date was not mentioned at the event.

New Alabama robotics research and development facility announced

The city of Decatur, Alabama has been selected as the site of a new $71 million dollar state of the art robotics learning center. After several years of looking at different places where to build the robotics hub, Decatur was finally revealed as the winner in an announcement given by state governor Bob Riley earlier today.

The facility will be dubbed the Advanced Technology Robotics Research and Development Complex and be spread across three buildings, each with a different focus in the field. The first building will be 40,000 square feet and be used to train the next generation of robotics engineers. Companies that specialize in the field of automated robotics will be invited to show off their wares and allow students to get practical hands on training. It’s estimated that this facility will be able to train about 450 people in the course of a given year.

The second building will specialize in the research and development for robotics used in space and military exploration. Both NASA and the U.S. Army Missile Command will be involved in the operation, testing and education of these developing fields of robotics.

The third building will be focused on the development of robots for emerging industrial applications. Students and companies will be able to test out new approaches to automated manufacturing techniques, train staff and test out new equipment and software.

At the announcement ceremony Gov. Riley said “Our vision from the beginning was a state-of-the-art robotics campus affiliated with a two-year college that is close to Huntsville’s aerospace research hub and easily accessible to our manufacturing industries. Calhoun Community College is a perfect fit for what we envisioned.”

Toshiba Satellite A305-S6845 notebook with face recognition security

Toshiba has unleashed another notebook computer in the Satellite series upon us and even though this blog is not interested in reviewing portable computers, the A305-S6845 has a feature we care about. The new Toshiba computer includes face recognition using the build-in web camera to allow owners speedy logins.

For starters, this is another notebook with all the bells and whistles that one would expect from a modern portable computer. It has a large screen and lots of computational power under the hood to make the machine I am typing this article on look like no faster than an Apple II.

Back to the notebook's face recognition capabilities, all I can find out about it from the product page is that the $1250 computer comes with the Toshiba Face Recognition software. Apparently, it is an additional security component supplementing the computer's integrated fingerprint system (which is actually a much more mature, reliable, and useful feature on this computer.)

Why this story is interesting enough to prompt me to write a post is that the Satellite A305-S6845 notebook is one of those low cost machines that many will buy. This effectively brings face recognition technology to the masses in an indirect way; Toshiba is not making a big fuzz about the computer's computer vision skills. And why should they considering that face recognition at this stage is hardly a reliable enough technology to secure valuable data. If you get one of these Toshiba notebooks, make sure not to grow a beard after you have trained the face recognition system or take off your glasses before trying to get access to your Vista partition.

If you don't have an A305-S6845 or never plan to buy one in the first place but you want to give face recognition technology a try with your computer, you should check out a free piece of software that we wrote about a few months ago Banana Security's LemonScreen.

Robots podcast replaces the Talking Robots podcast

The researchers from After a successful 2-year run with a total of 45 interviews, EPFL's Talking Robots podcast will be discontinued and replaced by the new Robots podcast. It looks like the people behind Talking Robots decided to move out from EPFL and invest in creating a more professional podcast (this is not to say that the original show was poor but any savvy listener could tell that it was produced by students.) In fact, the same people responsible for Talking Robots are also the people behind the new podcast.

At any rate, the new robotics podcast will also center around the latest news in robotics focusing on interviews with the people behind the scenes in research laboratories and the industry. There will be a new show published once every two weeks following the tradition of the original podcast.

The Robots podcast published their inaugural show on June 6th was focused on DARPA's 2007 Urban Challenge. The hosts interviewed Professor Daniel Huttenlocher from Team Cornell and Rick Yoder from Velodyne who have created a LIDAR sensor central to many Urban Challenge teams' designs.

I wish the producers of the Robots podcast the best of luck and much success with their project. I will be tuning in fortnightly to check out the latest show and so should you.

PS: If anything, you should at least check out final episode of Talking Robots dedicated to what the general public thinks about robots. Get it here.

Using software to predict the outcome of sporting events: Can it be done?

There are two questions both of which I am almost always asked when I tell people that I have a degree in Artificial Intelligence. One is whether I can create software to predict the stock market and the other whether I can write software to predict the outcome of sporting events. The reason why these questions are asked is the obvious one. If such software existed then it would make its owner an instant millionaire.

I decided to write this post in order to give my answer to the second of the two questions the one with regards to predicting the outcome of a sporting event. The 2008 Euro Cup tournament will start a few hours from me posting this article and for the last couple of weeks several people asked me if I knew of any software to predict the winner of this tournament. If you are wondering whether many people are interested in knowing the answer to this question then just search on Google for “Euro 2008 predictions” and you will get back over 2.5 million results. I guess many people want to improve their chances of winning when betting on the outcome of this soccer tournament.

So, can AI predict the results of the Euro Cup?

The direct answer is “No.” Even though at least one group has made an effort to predict the outcome of the Euro Cup, there are too many variables that are often left out making any predictions as good as any other guess. Maybe you can say it is an educated guess but it won't be any more educated than the one made by your local newspaper's sports editor. The group I am talking about has a long history of trying to predict the outcomes of sporting events and some times get it right. The Swiss bank UBS modified software used for economic predictions to give their own set of Euro 2008 predictions. The actual predictions seem ludicrous to anyone with some basic knowledge of European soccer; the chances of Portugal being eliminated in the first round of the competition and the Czech Republic winning it everything are not the choices an informed soccer enthusiast would make.

I said earlier that UBS has had some luck making predictions about sporting events. The truth is that 2 years ago they correctly predicted that Italy would win the 2006 German World Cup. Then again, many had predicted the same considering that Italy has been a major force in European and World soccer. However, we should give UBS the benefit of the doubt considering that the last European Cup in 2004 was captured by Greece, a country that had not won a single game in the Euro or World Cup before and in fact had not even scored a single goal.

So, what makes sporting events so difficult to predict? As I mentioned earlier, there are far too many variables to consider anywhere from which players will be available at game time in top form, what the weather conditions will be like during the game, what the relationships among the players of each team are like and so on. AI software usually looks at the results of the teams in the past and tries to guess how the same teams will perform this time around. But the teams are not the same anymore as players and coaches come and go and the overall talent in a team can vary greatly over time. No software used for predicting the outcome of sporting events takes such parameters into account. If they ever happen to make a correct guess, chances are that they got lucky and nothing more. At the end of the day, anyone who has heard of soccer can tell you that Italy or Germany have a better chances of winning the Euro Cup than Austria and Switzerland!

Virsona lets you talk to anyone

Imagine being able to talk to any historical figure from history from sitting at your keyboard. Albert Einstein, Cleopatra, Adolf Hitler, Jesus Christ, even Babe Ruth if you so wanted, could be chatted with. Take that idea a step further and suppose if you could talk to fictional characters as if they were real people. Do you want to ask Peter Parker what his day is really like having the secret superhero identity of Spider-Man to protect? Or what about asking Sherlock Holmes for a little advise about solving a mystery? If the technology behind Virsona holds up that is precisely what you may be getting the opportunity to do.

At the Digital Hollywood conference held this past week in New York City Peter Hodge, the CEO/co-founder of Virsona, was giving the world the head’s up about his company. “Virsona Inc. is a company that brings to market a virtual persona, an online digital persona, of a person or a historical or fictional character. It’s a real-time chat interface. It looks like you’re having an instant messager conversation,” explains Hodge.

Hodge claims that anyone can teach his software how to hold a conversation just like the real character or figure from history after spending a couple of weeks inputting information into the database. The data appears as diary-like entries on your browser screen. As you ask the Virsona personality questions it will respond using the knowledge and personality that has been programmed into it.

Virsona is hoping to attract the attention of Hollywood hoping that the lure of letting the prospective audience have the chance to “talk” with the characters from movies will generate bigger box office. The application could also be used for education where kids could surf onto the site and learn the basics about science from a Virsona persona constructed to resemble Isaac Newton. The doors officially open on the Virsona website next week.

Rescue robots hold potential for saving lives in disaster zones

When a natural or manmade disaster strikes the images we see on the television news can be heartbreaking. No one can forget the chaos created by Hurricane Katrina, the collapse of the World Trade Center towers, the Southeastern tsumani of 2005 or the recent earthquake in China. When rescue workers reach the rubble often they are hampered because their own personal safety is now in jeopardy and they wind up losing precious time until the area can be secured for human exploration. But if disaster agencies follow the lead of Robin Murphy and her robots then more lives could be saved from harm’s way quicker than ever before.

After seeing the destruction caused by the Oklahoma City bombing, Murphy had an idea to create robots that could crawl into small, hard-to-reach holes where often there was no visible light, to locate survivors or find easier routes for rescuers to use. Using her education as a mechanical engineer and computer scientist, and the assistance of the USF’s Center for Robot-Assisted Search and Rescue, she created one of the world’s first robots made specifically for disaster relief operation. Her robot was used to scour the debris of the World Trade Center, going deeper than any rescue animals could and in places where the rubble was still smoldering. Ever since she has been called in by state governments to assist in rescue recovery including the aftermath of Hurricane Katrina in New Orleans. When Murphy sees the devastation caused by the Chinese earthquake or the cyclone in Burma, she sees that countries all over the world need robots like hers to save lives.

She might get that chance. Microsoft has seen the potential in Murphy’s disaster relief robots and given her half-a-million dollars to create an improved version of her robot called the Survivor Buddy. These robots could crawl through the wreckage of buildings to locate survivors and give them emergency water and a way to communicate to rescuers. The robots could even be used on the battlefield to get to soldiers pinned down by enemy fire. Larger versions could serve as a shield allowing the soldier the means to get to safety.

"It's a relatively small grant, but it gives us a chance to jump-start these crazy ideas and this hardware," Murphy told the St. Petersberg Times. "This money is hard to come by because this idea of human-robot interaction has only recently been recognized as an important scientific field.

"My dream is that one day you'll see rescuers and dogs at a disaster site, but if you don't see a robot you'll say, 'Where are they?' because they'll have become so commonplace. They'll do things dogs and people can't."

Sparky the dog robot aims to build better understanding about your environment

Sparky may be a common name for a dog but the one designed by robot engineers is the only one of his kind. Made to be the world’s “first environment health awareness pet”, Sparky is a robot that, when it interacts with the public, gives them a better understanding about the environmental quality of life around them. The robot’s exterior is constructed from wood and plastic while inside Sparky there are several sensors, a motor and strings which are connected to the robot dog’s appendages. The sensors register the amount of carbon monoxide, cigarette smoke and alcohol in the air quality around Sparky, as well as the amount of lighting, the noise level and how hard Sparky is being touched. All of this data is fed to the team’s nearby computer where it’s interpreted and given a value ranging between -100 and +100. When conditions are negative around Sparky the robot will act negatively by coughing, whimpering or even collapsing down. When Sparky is happy about his environment it will make uplifting sounds and even musical notes. Giving him a “walk” by pulling the rope on his wood box will pick up his mood the fastest.

“One thing we are hoping to explore are the ways that Sparky can influence or make people reflect on awareness regarding physical health and how environmental factors can effect a smaller being (giving smoking and alcohol negative connotation),” it says on the project’s website. “It also acts as a negative projection tool, (people we have tested on aim to have a ‘happy’ dog, and one that is collapsed and whining almost seems embarrassing).”

So far the project team has taken Sparky to the streets of Malmö, Sweden, but they are hoping that interest will come from environmental organizations or businesses that want to promote ways to make healthier living. You can find out more about Sparky at http://sparky.workx.biz.

Vote for the best robot comic

NOTE: Robot comic poll on the sidebar to the right of this post.

The recently concluded International Conference on Robotics and Automation (ICRA 2008) included for the first time a very cool contest. Conference participants were asked to submit their entries for the first ever (and hopefully not last) robot comic competition with a first place prize of $500 and an IEEE-RAS membership for 1 year. 13 participants submitted their entries and a panel of 5 judges selected what they thought was the best one. I won't tell you which was the winning entry as of yet because I want to actually give the people of the Internet the chance to vote for what they think is the best one.

There is a poll on the sidebar to the right of this post. Here you can cast you very own vote for the best robot comic of ICRA 2008. The 13 entries are given below at no particular order (click on the descriptions to see the comics.)

1. Unsung Heroes of Robotics


2. Cognitive Robot


3. Mover-bot


4. Drawing Robot Destroys CAD City


5. Robot Fish


6. Mona Lisa


7. The Robot Channel


8. Human-Robot Interaction


9. P3Z and Student


10. Random Walk


11. Robogeek


12. Blurry Vision


13. Robot vs. Human Debating Contest

Credit for these comics goes, of course, to their creators. Some have signed their creations and others haven't. For those in the latter group, if you are reading this, then please post a comment to claim your comic. It is not my intention to claim credit for your work; it is that I could not find detailed information about the robot comic participants on the official robot comic website.

Part of the reason why I wanted to create this poll is the fact that I disagree with the winning entry chosen by the judges. However, I should point out that I think all the comics were excellent and all participants deserve congratulations for their talent and efforts. I am curious to see if a popular vote will agree with the judges or not. For the result to be meaningful, we must get as many people as possible voting, so pass this on to your friends and colleagues.

Learning robots: A new IEEE RAS technical committee

Starting this year and with a first meeting during the IROS conference to be held in Nice, France, this coming September, a new technical committee initially chaired by MIT's Nicholas Roy will come into existence. The new technical committee will focus on learning robots that is robots that can learn representations and action sequences from their interaction with their surrounding environment. This has long been a desirable goal for Artificial Intelligence and Robotics researchers.

There is an increasing interest in machine learning and statistics within the robotics community. At the same time, there has been a growth in the learning community in using robots as motivating applications for new algorithms and formalisms. Considerable evidence of this exists in the use of learning in high-profile competitions such as RoboCup and the DARPA Challenges, and the growing number of research programs funded by governments around the world. Additionally, the volume of research is increasing, as shown by the number of learning papers accepted to IROS and ICRA, and the corresponding number of learning sessions.

From looking at the learning robots technical committee website, many well known and widely successful robotics and machine learning people have already signed up to participate. Some familiar names include Sebastian Thrun, Michael Bowling, Nicholas Roy (as already mentioned,) Gordon Cheng, Aude Billard, Oliver Brock, Lydia Kavraki, Andrew Ng, and many others. This is definitely a group whose actions are worth following over the next few years.