A European Social Robotics project called CHRIS (Cooperative Human Robot Interaction Systems FP7 215805) has received its final review last april. They have also created a very nice video that summarizes their work:
As the video shows, the work is about safe cooperation and it includes the recognition of gesture (pointing), speech, actions, and objects. It is quite interesting to see the capabilities of a 18 month old human child compared to what a robot can do (not much). Perhaps that is due to the fact that kids never worry about hurting their parents. They just go for it, and their mother will let them know when they are doing something they should not. That is a pwerful learning strategy.
I think the CHRIS video shows nicely what Humanoid robots can and can not be expected to be able to do. Most robots will not be able to do everything the iCub or Bert2 can do. CHRIS is at the cutting edge.
Size. With robots, does size matter? One might say size matters little if it works well and performs the job it was designed to do. From an engineering perspective one might even add ‘the smaller the better’ or ‘the cheaper the better’. But from a psychological perspective it may be very important how big a robot is. In the social interaction with humans a big robot may function differently than a small one. Big impresses, big attracts attention, big frightens. Small endears, small puts at ease. Those are just some associations that we should think about studying closer.
An interesting ongoing experiment with size is the work of Hajime Sakamoto and his Hajime Research Institute (a Japanese company building humanoid robots). He plans to build larger and larger robots that can walk on two feet, eventually ending up with a ´Gundam´ robot that can seat a human (co)pilot, as featured in the famous SF-cartoons (of which Hajime is a fan). A fairly recent prototype can be seen in the following movie:
Another nice example of the effect of a big robot size is the fake robot Titan, see the movie below. He attracts a lot of attention, and scares people a little, but also puts them at ease quickly by using humour.
The very small robots, like Keepon, are perhaps relying partly on their smallness to endear people and to easily establish contact with people. An interesting thought experiment is to imagine a really big Keepon, say 1.50m, trying to interact with a five year old autistic boy. Do you think it will be just as succesful as the small one? I do not think so. Or imagine a 2m Paro robot ‘baby’ seal. Hmmm, that might actually have an effect similar to that of those giant teddy bears you can win at fairs. Holding them in your lap would be difficult though.
Here is an interesting new robotic surgery assistant: the ARTAS™ System. It was recently cleared by the FDA (read more). ARTAS apparently helps with something called ‘hair follicle harvesting’, according to Restoration Robotics, the company that invented and produces the system.
The ARTAS System (Source: Restoration Robotics, Inc.)
The procedure is as follows. The client first sits in the Artas chair, and then his hair is millimetered. Then, a robotic arm equipped with a camera initiates ‘small dermal punches’ and harvests individual follicles. This is under the control of a doctor. The follicles, which are later transplanted by hand, will start producing their own hair over months.
Here is a paper (PDF) in the Dermatology Times that reports the results of trying out a prototype of the device. Apparently, no sutures or bandages were required and using ARTAS is quicker and less invasive than other hair transplantation techniques, like strip harvesting where a strip of skin with hair is transplanted to a balding area. The company expects to reach extraction rates to 750 to 1,000 follicular units per hour. In addition, it may require fewer staff (although robot support engineers should probably be on standby).
A very interesting, ‘social robotics’ study recently gave us all some nice resultats. Apparently people do not always appreciate being touched by a robot.
Somewhat surprisingly the results from one of the experiments showed that it matters why the robot touches you. If people think (because they are told) that it is because they will be washed then that is okay, but if the robot touched someone to comfort them then they found it much less agreeable, even though the touch movement was exactly the same. Apparently instrumental touching is more acceptable than social touching. And the perceived intention is what matters, according to one the researchers, Charlie Kemp. However, if you compare this result to the positive responses generally reported with the huggable robot Paro, then I think that this result may depend to a large extent on the actual appearance and exact behaviour of the robot. In this case the apperance and behaviour of the robot, Cody, may have created a mismatch with an intention to provide a comforting touch. In other words, the robot does not look like or act like it is designed to provide a comforting touch, it looks like it is designed to clean people (which is exactly what it was designed for).
In addition, the results showed that people did not like it if the robot announced that it was going to touch them, perhaps, as indicated by the researchers, because the voice startled them. Here, I think it is very important how a robot speaks exactly. If it speaks with a moving mouth and facial gestures, then this comes across as if the voice is coming from the robot. If a robot has a face and mouth that are able to ‘speak’ then people may actually expect a voice. But, if a robot speaks ‘out of nowhere’, for example if it merely plays a soundbite through a speaker, then this can easily startle people. It is a disembodied voice. So, again I think that follow-up experiments should be done to provide more conclusive results (as also suggested by the researchers). In a way this resembles the previous critical remark: the robot does not look like it was designed to talk to people so it may come across as a mismatch if it does talk.
But, all in all, this sort of research is very useful and more of it is needed to support the succesful introduction of healthcare robotics.
De second International Conference on Social Robotics, ICSR 2011, will be held in Amsterdam on November 24-25. The Call for Papers is out (the CFP page, and here the Full CFP Details).
De tijdslijn is als volgt:
Paper submission: June 1st
Notification of acceptance: August 1st
Final manuscript submission: September 7th
Conference: November 24-25
The International Conference on Social Robotics brings researchers and practitioners together to report on and discuss the state-of-the-art research in the field of social robotics. The conference focuses particularly on social interaction between humans and robots, the integration of robots into our society, and the design of next generation social robot interfaces and systems.
The theme of the 2011 conference is “Alive!” It expresses the vitality of the social robotics research, paying particular attention to the development of robots that appear increasingly social — the point that people perceive them to be alive. The conference aims to foster discussion on the development of computational models, robotic embodiments, and behavior that enable robots to act socially and the impact that social robots have on people and their social and physical environment.
Ben Robins, a researcher who has already done a lot of work studying how robots might benefit children with autism, is quoted as saying:
“Children with autism don’t react well to people because they don’t understand facial expressions,” said Ben Robins, a senior research fellow in computer science at the University of Hertfordshire who specializes in working with autistic children. “Robots are much safer for them because there’s less for them to interpret and they are very predictable.”
The article neatly decribes the current state of the science behind the idea that social robots can help autistic children to learn and train certain social skills (basically, there are promising case studies, but a long-term effect study is lacking). And it mentions the opinions of various researchers, in the field and outside of it, on the merits of the work with Kaspar (which has been ongoing since 2005).
I also found a nice BBC video from 2008 about Kaspar and Robins and others’ work:
And there is a long, Japanese documentary about Kaspar and the work of Robins et al.
Recently, on October 25, Jeroen spoke at a workshop about healthcare robotics. It was organised bij Kennisalliantie and Syntens, who wish to set things in motion, especially in the Dutch ‘Medical Delta’ (roughly Rotterdam-Delft-Leiden). Prof. Luc de Witte opened the day, followed by Boudewijn Wisse, and finally Jeroen Arendsen. In the afternoon the discussion was continued in groups. The video gives a good impression of the day. For Robots that Care the initial contact with Zorgbelang, represented by mrs. Aat Hoffius, may well lead to further developments.
ScienceDaily (Feb. 3, 2011) — Surgeons of the future might use a system that recognizes hand gestures as commands to control a robotic scrub nurse or tell a computer to display medical images of the patient during an operation.
Purdue industrial engineering graduate student Mithun Jacob uses a prototype robotic scrub nurse with graduate student Yu-Ting Li. Researchers are developing a system that recognizes hand gestures to control the robot or tell a computer to display medical images of the patient during an operation. (Credit: Purdue University photo/Mark Simons)
I believe, partly based the insightful comments given at Nursing Advocacy, that it is not a good idea to replace human scrub nurses with robotic ones. It would be much better to solve the shortage of nurses). In this case, there is so much a scrub nurse does, so much more than a robot can do, for at least the next fifty years or so.
But there will be a publication in the Communications of the ACM soon. That will be an interesting read
Update: There is already a nice video about the story on the ACM site (here).