It’s not so long from Christmas and I just remembered that I wanted a robot for my seventh Christmas. That was in previous millennia, so it was not an easy task for my parents when I wrote all the features I wanted. I got one but was quickly disappointed by it. The tech and the budget were obviously not there yet.
When I fast forward to today, I can see consumer robots are becoming much more than just toys and smart vacuum cleaners. Advancements in the – so called artificial intelligence – spans normally also to robotics, and we are certainly living in a world that will soon see humanoid or android robots as a part of our households. Previously a matter for science fiction once again coming to reality.
Some robots are trained as personal assistants, humanoid service robots are making sure the kitchen will be clean at all times and even cook and deliver the meals, taking care of garbage and getting to all areas that we as humans want to “outsource”.
But it goes beyond that – some robots are already security specialists and guards and some are even evolving to be our emotional companions and try to cross the border to be our virtual friends.
People are able to clone voices and even memories of real people can be integrated into robots, planning to make them companions for people struggling with complicated dementia and some mental health issues. Healthcare sector has already had pilots for different use cases in hospitals, ranging from basic care, physiotherapy to surgery. And a single search in your favorite search engine will probably find thousands of other applications of using robots for care, seems like imagination is becoming the only limit.
What does robots have to do with accessibility?
Thanks for asking. It may not be obvious, but robots need to have interfaces as well. We can all imagine just talking to them, but that is in fact not really accessible. We may use brain-computer interfaces (BCI) quite soon, but what if we don’t want to. Or if they are not compatible with us. Or if that we find out that they are dangerous and so on? Then we still need to think about more classical human to robot interfaces.
Some robots that I have actually had the pleasure to see personally rely on application programming interfaces (API), so they can be controlled with software on your device, like application on mobile phone or even web application. This is then logical as it comes to accessibility – we need to make that interface accessible.
But what when we consider other interfaces?
Sound based interfaces – voice control, robot speech
I already mentioned one of the most popular ones – voice control. It would feel, in a way, like we talk to another human, especially if the robot can also speak back to us (and that is fairly easy with technology available today). But if this voice control will only be trained with limited training data it will be a problem for many groups of people. Dialects, temporary voice problems (like laryngitis (inflammation of the voice box in our body) would create huge barriers or even make it impossible to communicate with them.
Not to even start thinking about people that can’t speak, or that doesn’t hear that good, or even doesn’t hear at all. How will this kind of robot communicate then?
Robot – human interfaces need to account for people with disabilities and we need to act quickly and apply accessibility to such interfaces otherwise we will leave people out, and with time we will leave ourselves out.
Just a quick reflection about why accessibility matters and why we need to act quickly.
Visual interfaces – gestures and screens
I think we understand that voice interface alone is not enough. And robots talking back is also not enough. We also need the interfaces to work for people that can’t talk and can’t hear. That is the first step to make robots more accessible.
Gestures may work, but only if people are proficient in them. Sign language is an excellent alternative to vocal language, but it does have the limitation of clear line of sight and it is basically much more complex than just using hands, so once again not very accessible for people not proficient in it.
Screens are excellent visual interfaces, but in the terms of robots they don’t quite do us a favor. Having to go to the robot, or having it to come to you to be able to communicate is kind of degrading the user experience. And taking time. And even not possible in some situations. So screens need to be remote. And then we come back to our smart watch, smart phone, laptop, tablet, PC or your virtual reality goggles.
Or any future device with screen. That could be doable. Once again if the user interface on that screen is accessible. Virtual reality googles are not very accessible for blind and deaf blind people, but we can make other screen based devices very accessible. Not perfect, but possibly the most accessible of all interfaces considering large and segmented group of people using them.
I think that classical web applications or native mobile applications would actually be possibly the best and most accessible way to communicate with robots considering all possible disabilities we people may have. Provided, again, app or web app is accessible.
One of many advantages of using a mobile or web app for communication is ability of using familiar assistive technologies with it. So in reality smart watch / smart phone / PC and app running on it is just a layer between the robot and assistive technology that people with disabilities may use.
Further resources on robot accessibility
Web Content Accessibility Guidelines (WCAG 2.x) try to be technology agnostic as much as possible (even if web is in their name) and are currently the most widely accepted standard for digital accessibility, focusing especially in wide-spread interfaces like web, mobile and documents. Virtual Reality (VR), Augmented Reality(AR), eXtended Reality (XR) and all interfaces beyond that are not represented directly, but it is possible to use some parts as guidelines for robot interfaces as well. Depending on actual hardware details, at least.
W3c Accessibility Guidelines (WCAG version 3) are still an early draft and I guess also a perfect candidate for standardizing the accessibility of robots. But realistically, we can’t really expect them to add a lot for robots specifically as there i still a lot that needs to be defined for native mobile apps and VR, AR and XR. So if you have hands on knowledge and possibilities to contribute I would strongly suggest doing it under WCAG 3 (opens in new window).
When I wanted to do a quick research on academic articles connected to robot accessibility I wasn’t able to find a lot, but I will still provide some leads here, just some of the articles that were interesting to me.
- A Proposal of Accessibility Guidelines for Human-Robot Interaction (opens in new window) – seems to be well structured and includes a lot of potentials, but it seems a bit outdated in parts. Which makes sense when we consider how fast the technology is evolving. I also liked some references to WCAG, as they are crucial for accessibility of known interfaces.
- Accessibility Guidelines for Tactile Displays in Human-Robot Interaction. A Comparative Study and Proposal (opens in new window) – although limited to tactile displays it offers also some extended proposals for additional guidelines that may help making other types of interfaces more accessible. Also partially references to WCAG and some other guidelines.
- Performing Accessibility Evaluations for Social Robot (opens in new window) provides WCAG-EM based Accessible Socially Assistive Robots Evaluation Method (ASAR-EM) that was used to evaluate some common social robots. Interested facts, but it also shows that such research needs to be regular, to cover all the hardware and software advances made. I liked the WCAG-EM based approach that confirms my thesis that we need to update existing standards and not create them from scratch.
- Accessibility Requirements for Human-Robot Interaction for Socially Assistive Robots (non-tagged PDF, opens in new window) – unfortunately I was not able to find a tagged PDF, so this one may be inaccessible for people with disabilities, which is not optimal considering valuable efforts in the content. Research is limited to Social Assistive Robots (SAR) and I liked the practical research done on the robots with people with disabilities.
- Accessible Human-Robot Interaction for Telepresence Robots: A Case Study (opens in new window) – researching augmented reality user interface for social telepresence robot with useful info. Unfortunately they decided not to test with people that are blind, has severe cognitive challenges and some other groups.
Appeal to all robot vendors and researchers
Please, please don’t forget about people with disabilities when developing interfaces for robots. Consider that this also means people in situations that disable them, it’s a complex world out there and if you make the interface to your robot accessible it will help much more people than between 15 – 20 % of population.
Please also train your robots to work with and around people with different disabilities. It will make your robots better for everybody and hey, you will be able to sell them to places and people that could potentially benefit the most from them.
Although legislation is known to be a bit behind technology advancements, especially in these times of amazing technological progress, remember that as soon as robots will come to our homes we will get laws to protect people. And protecting people with disabilities will be, I really hope, one of the top priorities.
We also need more research on the intersection of accessibility and human to robot and robot to human interfaces. The earlier you start, the better. Billions are invested in sensors, improvements, software, learning, so there should be enough of resources to also make robots accessible. Lack of awareness must not make the world full of barriers for people with disabilities, again. We need to take action now!
I’ve seen some studies on the accessibility of humanoid robots, but we need way more research to be able to prevent discrimination of people with disabilities. Please include people with disabilities and accessibility specialists or otherwise you can choose wrong paths and the result will discriminate somebody that is probably already struggling with the barriers all over the place. It’s like with other digital products – starting early is the only sustainable way to make anything accessible, and it goes for robots as well.
