Will The Next 10 Years Bring Humanoid Robots Into Our Homes?

The holy grail in robotics has long been the humanoid robot that can live and function in the home, something similar to what viewers of science fiction movies have grown accustomed to seeing — the robot greets you at the door, takes your coat and then puts away the shopping before getting dinner ready.

In the dystopian variant — think Isaac Asimov’s I, Robot, which went on to become a blockbuster film in 2004 with Will Smith — something goes horribly wrong as the robots begin to think for themselves and rebel, forcing the (human) hero to save the day. Notwithstanding stories about robots that would rather dominate humans than help them, robotics technology has progressed rapidly in recent years, with the conversation shifting to the question of when we will have humanoid robots in the home rather than if it will happen.

“The revolution will begin in 2014,” predicts Bruno Maisonnier, CEO and founder of France’s Aldebaran Robotics, a scheduled speaker at LeWeb. “Many people think an affordable humanoid robot that can do useful things is not close, maybe in ten years they say. I’m happy all those people are not entering this field.”

Maisonnier’s bullishness is not shared by everybody in the industry. One of those who say much more time is needed is Limor Schweitzer, who founded RoboSavvy in 2004 and has made a name for his London-based company thanks to a small prototype humanoid dancing robot named Fonzie. Schweitzer says Fonzie proved it could be done, but going from there to building a large humanoid robot that can do useful functions in the home and won’t require a large mortgage to buy is still a way off.

“As you increase the size of a robot the cost goes up exponentially because it has to sustain its own weight,” says Schweitzer. “They can be dangeros due to their size, weight, power and experimental nature. For €25,000 you might be able to have a robot that brings you a drink on demand, but I wouldn’t necessarily sell it to you unless you are a hobbyist or a researcher because it wouldn’t be a stable system. You can definitely get a robot to do some cool stuff, but it would be cheaper to hire a low-cost employee who could do it for you.”

Honda, the Japanese car manufacturer, has developed Asimo, a humanoid robot that already fills some of the roles one might expect in the home such as getting a drink when asked, but its cost makes it unfit for the consumer market.

“The Honda robot is terrific, but it costs $1 million and you can’t have robots in the home at that price,” says Maisonnier, winner in 2012 of the Marius Lavet Prize, which is awarded yearly to a French person whose innovation is acknowledged as a technical and commercial breakthrough.

To contain costs, Maisonnier is keying on creating a robot that can have meaningful interactions with humans, including understanding the nuances of human body language. He says he will worry later about making sure that a robot can cook and do the laundry.

“We need to have a robot you want to have at home, that you accept, that you are ready to interact with,” says Maisonnier. “Too many people are focused only on usage and utility. I’m not in agreement with this trend. If you want a robot it’s not about left-brain usage (logic and analytics). What’s important in life is dialogue and emotion. Our goal is a robot that is a companion and partner, that is kind to you, that loves you.”

Aldebaran, founded in 2005, has sold more than 5,000 of its first product, a 58-cm-tall humanoid robot called Nao (pronounced “now”) which is being used in some 500 universities and research labs including Harvard and Brown in the U.S., Ecole Nationale Supérieure des Arts et Métiers in France, KAIST in Korea, and University of Tsukuba in Japan. The company has 350 employees in France, Boston, Shanghai and Tokyo and had sales of €8 million in 2011, €18 million last year and the forecast is for €35 million this year.

While Aldebaran puts the finishing touches on a version of Nao for the general public, the robot, which can see, hear, speak, communicate and walk, is already being used to help children with autism break down the communication barriers that often make it difficult, if not impossible, for them to interact with people. Researchers have found that autistic kids often have an intrinsic interest in technology and can interact with a robot in a way that is initially impossible for them to do with people. The idea is to use the robot to teach autistic children how to read and react to the confusing social and emotional cues that people without autism take for granted.

In a sign of the importance Maisonnier puts on the human interaction side of his humanoid robot plans, in October Aldebaran said it would begin integrating into Nao voice recognition and text-to-speech technology developed by Massachusetts-based Nuance Communications. Aldebaran, which was already using an earlier version of the American company’s voice recognition technology, says the new agreement will let people have a more natural conversation with robots in 19 languages. Nao robots with the new Nuance technology will ship early next year.

“Developing natural, conversational interactions with humanoid robots is a challenging and pioneering area of robotics,” Maisonnier said when the agreement with Nuance was announced. “Our robots have to express and understand emotions and for this there needs to be expression in their voice that speaks to their personality, and matches their body language.”

Nao costs about $15,000 though Maisonnier expects Aldebaran will soon be selling humanoid robots that costs less than a third of that price.

Aldebaran is leading a project to build a bigger robot called Romeo, which will be about 1.5 meters tall and is designed to help people who are not completely autonomous. Almost half of the €10 million budget to develop Romeo, the first of which will ship earlier next year, has been subsidized by the French government with several other companies working on the project as well. Maisonnier sees Nao and Romeo as intermediate steps on the way to realizing a humanoid robot for the home.

Aldebaran has made a massive push forward in the past year thanks to an influx of funds from Japan’s SoftBank, which bought 75% of Aldebaran in 2012 from pre-existing investors including Intel Capital and iSource for a reported $100 million. That allowed Maisonnier to hire more people to ramp up production and accelerate development.

Maisonnier concedes there are difficulties in reaching the promised land of the affordable humanoid robot for the consumer market, the most arduous of which is integrating seamlessly the more than 200 technologies that deal with everything from vision, speech, movement, balance and touch. “If just one doesn’t work you won’t have a positive experience with the robot,” he says.

RoboSavvy has instead decided to concentrate on the entertainment and research side of robotics. Schweitzer says that for the foreseeable future, to be financially sustainable a robot has to meet the needs of one of those two markets. RoboSavvy is also meshing robotics with 3D printing, its other line of business, in a bid to reduce production and maintenance costs.

“The plan is to take advanced robotics away from the Ivy League of MIT and the labs that have all the money and bring it to the younger generations, to schools and to developing countries where it would not be possible due to cost,” says Schweitzer.

While the structural parts can be printed, electronics and the motors have to be acquired. He says the total cost of the material to produce a small humanoid robot through 3D printing would be about $1,000. Getting from that small humanoid robot to a large one that can function effectively as a servant in the home is still far off, Schweitzer says.

But that doesn’t mean we won’t be friending robots in the next 10 years. RoboSavvy recently developed a small robot that can interact with humans by gleaning information from Facebook. The robot asks the person to become his friend on Facebook after which he plays music the person likes and makes comments based on information in the profile. It can be as innocuous as a complimentary remark about the city you live in or as personal as, “I’m sorry you are no longer in a relationship, but you are invited to a party at my place on Friday night.”

If Aldebaran and RoboSavvy are right, it’s a safe bet that in the future many of us will not only be interacting with humanoid robots, but “liking” them as well.