tl; dr: When we solve the problems of their communication with our environment, other devices and with us.

This article is a look at the situation with domestic robots from the beginning of robotics and the interaction designer. It will briefly highlight the most famous models of home robots and the expectations of their users, as well as examine in detail the reasons why these expectations do not quite coincide with reality. So...

What we have now

Over the past year, a record number of home robots has appeared on the market. Jibo , a small but very expressive ball on the stem; Buddy - modular home robot with removable multi-functional hands; Asus Zenbo - a robot capable of independent movement, positioned as an assistant in many household chores; and other similar devices. It seems that the next technical revolution is about to come, and behind it is a rebellion of cars and a doomsday . But let's take a closer look at these new robots. What is so new and special about them?

What can they do?

The already mentioned Zenbo and Buddy declare a rather similar list of functions. They can take and play photos and videos, make video calls and conduct video surveillance of the house. Like Jibo , they know how to keep a to-do list and reminders and read out loud important information like your favorite pasta recipe. They are also able to perceive voice commands. Some have a panic button. Interestingly, all these robots have educational and entertaining functions for children, such as learning, dancing to music and storytelling.

Home Buddy Robot is trying hard to look friendly.

It is easy to see that, with the exception of a few functions like a dancer, all these things can be done with the help of a regular tablet and, possibly, a home surveillance camera. In fact, these new home robots even look like tablets on wheels. But would we really see such home robots?

What would we like

Let's turn on the fantasy. If you could get any home robot you want, with any feature set for just 9.99 What would this robot be? What would he do for you?

The studies ( 1 , 2 , 3 ) conducted on this issue basically show the same picture: Most of us prefer home robots to do the dirty, boring, and non-creative housework for us. Cleaning floors, windows, walls, tables and shelves, washing dishes, caring for a garden - all this people would gladly entrust to robots. Less respondents already want to delegate creative tasks like cooking. But communication with children, pets and elderly relatives, education, reading aloud and other social tasks, people do not want to entrust the robots categorically.

What is the result? The robots we see on the market today do not do the things that we would like. Moreover, some of them, like Buddy and Zenbo , are stepping onto the slippery slope , taking on some of the social interactions that people generally would not want to delegate to them.
It is not difficult to predict that such devices are unlikely to make a new technological revolution.

What prevents us from creating really useful robots that will keep our house clean and not interfere in our relationship, as we would like to?

What prevents?

From my designer bell tower, there are three main obstacles to creating useful home robots. Here they are, in order of decreasing severity of the problem: the infrastructure that is sharpened for people lack of interoperability; unprocessed user interaction. If you see other or additional problems from your bell tower, please share them in the comments. So, more about each obstacle for home robots:

1. Infrastructure

One of the main obstacles in the path of assistant robots is our human way of life. The fact is that while our homes are not suitable for anyone except healthy adults. All the objects around us are made in such a way as to be distinguishable with the help of healthy human eyes, to comfortably lie in a medium-sized human hand and not to interfere too much with the movement of a person of average height with healthy legs. If you are forced to use a stroller for movement, then for a comfortable existence you will have to make several significant changes in the "standard" housing. If you are a one-year-old child, then your parents will have to buy you special furniture and utensils, as well as hide many dangerous items from you. If you are a domestic cat, then you will not be able to use for the intended purpose almost any human object, even if you know how to do it. If you are a home cleaning robot ... Then you must perceive the information, process it and act just like a person, otherwise you will not be able to do anything useful in human dwelling.

For example, take such a simple operation as wiping dust off the table. This routine action in human performance involves searching for a cloth, cleaning agent, possibly a bucket into which you need to get water. All these items need to be found in various boxes and boxes, located in arbitrary places at home, they need to be recognized, transferred to the table (which also needs to be detected and visually measured), then apply the cleaning fluid to the cloth, gently walk the cloth around the table, wash the cloth ( you already understand what the problem is?), wipe the table dry. If water is contaminated, it needs to be changed. After cleaning the table, the water should be poured out, and the objects should be returned to the places, if possible without dropping anything else. All this includes, at a minimum, the knowledge that there is a bucket, a rag, a bottle with a means (which may change from time to time), how to properly hold and use them, where to find them and so on. People do it without thinking. And robots? Well, they can already bring you a beer . Provided that you are sitting in the appointed place, and the beer is neatly placed on the arranged surface, free of mess other items. Creating robots that perform routine tasks similarly to humans can be very difficult, time consuming and expensive. And the obvious overkill.

Maybe robots still do not imitate adults? In the end, we buy this baby utensil for a small child and this cat toilet for Fluffy. Maybe we will invent something similar for a home robot? Special box with all the necessary rags, liquids and tools, for example? Alas, but as soon as you start to figure out in your mind how much you need to put in this box for your home robot to be truly universal, you will soon discover that the set is very large. In addition, the robot will still have to navigate in our home, safely move disturbing objects of different weights and stiffness, open doors equipped with different types of handles and opening in different directions ...
Well, in order to give the robot the opportunity to cook with human knives, pots and food, you will need to rebuild almost the entire kitchen.

Well, maybe we don’t need one robot doing everything at once? Maybe we should start with smarter machines to perform certain routine actions? While it looks like the most promising approach. We already have automatic washing and drying machines, dishwashing machines, smart robot vacuum cleaners and a lot of automatic polivalki of all shapes and colors. If we could automate some more difficult and unpleasant tasks, like washing plumbing, then we could have done without a universal cleaning robot. True, then you have to spend a little more time on turning on, turning off, maintaining and setting up all our smart devices. It would be nice to be able to work with them somehow centrally ...
And here we have a second problem.

2. Interoperability

This fashionable English word can be roughly translated as "the ability of different devices to interact productively with each other to perform common tasks." In fact, this means that in order to achieve the best result, the entire zoo of smart home devices that exists now and will appear on the market in the future should be able to communicate with each other. And, perhaps, with a home robot, which can serve as a “butler” and transfer your useful instructions to smart devices.

Why do you need it?

Here are just a few possible uses in which Interoperbility is needed:

• A robo cart that delivers the purchased food to the smart refrigerator that ordered the food. The cart should be able to interact with the refrigerator, with the door and with the delivery drones. Yes, I hope that we will have delivery drones.

  
  

• The ability to view the status, needs and problems of all devices on one smartphone, or with the help of a robot butler. Accordingly, a smartphone, a robot and the entire home zoo of devices must be able to exchange information.

  
  
• Your robot vacuum cleaner, a smart watering can and an automatic toy for a cat should not collide in the middle of your living room and interfere with each other’s tasks. To do this, they need to pass each other information about their location and, possibly, the planned path. Well, or be able to automatically avoid each other.

Alas, as long as business with communication leaves much to be desired. As far as the author knows, at the time of this writing, most manufacturers of "smart home" do not support a single standard for communication. At best, you will get an integration with IFTTT . In principle, at the very beginning this may be enough. Moreover, the interest of IFTTT is quite friendly to ordinary users. In the next paragraph it becomes clear why this friendliness is important.

3. User interaction

Suppose we created the perfect robot-friendly infrastructure and taught all our smart devices to interact with each other. Now it would be great if the user could manage all this wealth, and receive from him high-quality and timely feedback on the state of the dwelling. And this task is not as simple as it seems.

How would you like to interact with your home robot? As a rule, the easiest and fastest way of communication available to people is oral speech. In most popular films with the participation of robots, this is how people interact with them. And with this, it seems, there should not be any special problems, right? In the end, we have Siri and Google Now, who seem to understand us well. With a fast internet connection, of course. Alas, voice communication has a number of limitations.
First, speech recognition algorithms are still very imperfect, as can be seen by looking at a viral video about the shopping list and toothpaste. Secondly, this way of interacting with the robot is not suitable for people who, due to any physical limitations, cannot use speech. Thirdly, it may be necessary for the user to command something to the robot while physically away from it. In other words, only the voice interface is not enough.
The user needs to provide an alternative interface, preferably just as intuitive.

The less obvious problem of user interaction is the huge number of potential features and capabilities that can be provided by various devices (let us remember about Interoperability). All these features must be adequately represented by user interest. With an increase in the number of possibilities, the interface is usually complicated, which is undesirable for a wide range of users. Above the interface problems for robots, active work is currently underway, and some solutions turn out to be very non-standard .

In addition to problems in the design of interfaces, there is another important problem in the interaction with the user. Namely...

Ethics and Security

The problems of ethics and security are quite complex, if only because many moral dilemmas, like the problem of a trolley , have not yet been completely solved by the people themselves. Moreover, the preferred solution to ethical issues may depend on the culture in which the question is posed. But even if we abstract from high matters and difficult moral choices, we will have one very practical, very serious problem. Her name is security.

Home robots will be able to have access to a huge amount of personal data of their owners. Take, for example, Buddy , which can stream videos from your home. Or imagine a smart heater that boils water when you actually need it. If the government, corporations or hackers get access to them, they can literally find out when you are taking a bath! However, for some of us, the surveillance of the powerful may not be as important as the potentially all-knowing partners or cohabitants. Let us recall at least the recent scandal over the Barbie doll , which could record the voice of a child playing with it and pass it directly to parents. We will have to work hard on addressing these issues and adopting relevant standards before letting in an army of robots into our homes. But there is also a more serious threat.

One of the characteristics of robots is their ability to influence the environment. So far we have considered only useful influences. But if, for example, an attacker gains control of a nailing robot, he will be able to do a lot of damage and cause serious damage in the whole house. This is not some kind of innocuous screen-blocker, but quite real, physical damage caused to objects or even living beings. Accordingly, it is necessary to make sure that even if the home robot gets out of control of the owners, it will not be able to cause any harm. A large red switch in a prominent place, too, may not be redundant.

Total

Before we can create and use truly useful and useful home robots, we need to solve a few problems. Some of them are quite serious: ethical issues, safety, and the structure of a dwelling that has been honed for adult healthy people. Other problems are currently in the process of being resolved and can be solved in the foreseeable future.

We, the geeks, will have to work a lot to solve these problems. But the opportunity to go into your crystal-clear house to order the robot butler “to cook something delicious” and proceed to the bath already heated for you - it's worth it.