• aditya mehrotra.

what will the future look like? -abridged version- [opinion]

The following is an un-edited set of excerpts from a research proposal written to the MIT Media Lab City Science Group in August 2019. The contents of this may not be used for research purposes. It is a pure opinion piece.


background & introduction

If you take a quick second and Google the phrase “most advanced robots for the home,” or “most common home robots today,” what comes up is a wide variety of the same exact thing. Google Home, Amazon Echo, Apple Homepod (all considered robots) are working to automate the home via an intelligent voice assistant. Kuri (now no longer produced) and Temi are robots designed to interact with you in your daily lives. And iRobot and Neato are both working on making our floors a little cleaner. 


Now, these are all incredibly impressive technologies considering the fact that if you had asked a similar question a few years ago, you’d be hard pressed to find any example of a real home robot of any description, let alone one that was doing anything material at all (excluding kits like Lego Mindstorms and Arduino). But there’s still a problem, best articulated by none other than Bill Gates, and I quote.


“Imagine being present at the birth of a new industry. It is an industry based on groundbreaking new technologies, wherein a handful of well-established corporations sell highly specialized devices for business use and a fast-growing number of start-up companies produce innovative toys, gadgets for hobbyists and other interesting niche products. But it is also a highly fragmented industry with few common standards or platforms. Projects are complex, progress is slow, and practical applications are relatively rare. In fact, for all the excitement and promise, no one can say with any certainty when--or even if--this industry will achieve critical mass. If it does, though, it may well change the world…

… what I really have in mind is... the emergence of the robotics industry, which is developing in much the same way that the computer business did 30 years ago. Think of the manufacturing robots currently used on automobile assembly lines as the equivalent of yesterday's mainframes. The industry's niche products include robotic arms that perform surgery... and domestic robots that vacuum the floor…”


- Bill Gates, Scientific American


The point Gates is making is that, yes, there are robots in our homes and in the industry today, but they’re very specialized and haven’t been developed to their full potential. Voice assistants are extremely good at understanding users but are not really helpful when it comes to aiding them in physical tasks. Cleaning robots are extremely good at what they do, but the problem is cleaning is all they do, they have but only one application. Even newer, emerging robots simply don’t do enough or aren’t practical. Extending Gates’ parallels with the computer industry, robots these days aren’t like today’s computers. A laptop has thousands of possible functions from checking email to running engineering design software. They can be used for work or for streaming videos when you’re bored. They can be used for catching up with loved ones or typing out a document like this one. On the other hand, the robots in homes today, heavily resemble the calculators of the past. They can do one thing really well, but they can do only one thing. 


We’re spending so much time on technologies such as stereo-visual odometry, GPU accelerated computing, bio-inspired actuators, mock-skin sensors and other amazing components (which are absolutely necessary to the development of the industry) that are supposed to operate as a part of a larger system which, on most accounts, isn’t currently happening. We have a collection of amazing “features” and no platform product to put them in. As an aside, we have also barely started thinking about how to get a human to interact with a robot, to build trust.


This is the dilemma I thought of when I heard the question City Science has been working to answer, “What will the future look like?” To me, (after thinking about this state of the robotics industry for over a year, trying to understand the industry, and trying to generate ideas) imagining the future of robotics in the home perfectly falls under the many answers to this question. Robots in our homes can have such an impact on the future, if we can only imagine the possibilities.


the idea.

As is likely already clear, I propose designing a robotics system for the future home. There are many components of robotics that need to be developed going into the future of the industry - there’s no set robotics computer or hardware platform that really does what we want a hardware system to do, no standard and reliable set of sensors, and no real operating system developers can build and run apps on. These are all true problems, but there is research working on them already, and they will not fully solve the problem of the utility of the system as a whole. I propose a project that focuses on developing the concept of what a home robot means, what it can be, and how a human would interact with one. I believe the best way to fully describe my proposal is to list what this project is, and what it is not.

This research project is NOT about developing new computer science to help robots navigate around the home or understand their surroundings.


It is NOT about developing new sensor technology or new electronic systems capable of allowing robots to read and respond to their environments.


It is NOT about designing complex mechanical systems able to dexterously grip small objects, or do the most specific of chores.


It is NOT about creating the perfect robot operating system software or platform for apps and development.


And it is NOT about combining every robot out there right now into a large mess of a system that does everything at once.


It IS, in fact, an exercise in design and in imagining the future. This research is about answering the question: what will the future of robots in our home look like, what will they be able to do, how big will they be, how will they interact with our homes, and how will we interact with them? It’s about understanding the “features” the robotics industry has developed and designing a system that can use them together to create a truly useful home robot of the future. What this system will look like, what it will do, these are the questions to be answered through research and understanding. Once a clearer picture is developed on the desired functions and design of the system - sensors, computer hardware (such as GPUs), power systems, control and perception algorithms, mechanical designs (to an extent), that already exist will be combined into a platform that uses them to their potential. In that sense, it is about adding the “features” the industry has currently developed, into a “product” of sorts.


Why? Because the future of robotics is NOT in CS and it is NOT in mechanical engineering, it is NOT in power technologies or nano computing, it is in design. It is in design because the technologies now will enable us to make great robots, but it is imagination, creativity, and design that will build the true “future robot” by implementing these technologies. 

I fully understand  that this is no short-term research project, but I am committed to seeing it through because the future of robotics in the home is something I truly and deeply believe in. 


guiding questions developed.

In the previous year of research about the industry, and in thinking about designs for prototypes of an ideal home robotics system, I decided it would be good to gauge what people would like to see out of a future home robot. I put together a very simple Google form, and circulated it through Facebook, hometown connections, and MIT friends. The form contained some basic questions such as, “Do you own a robot? Which of these tasks would you like your ideal robot (if you owned one) to perform? What tasks not included here would it perform? Do you have any reservations to owning a robot?”


This form was by no means a survey with data that can be used to develop a system, it was by no means a carefully conducted research report on what people want out of their robots. It was an initial idea generation technique, a starting point. The responses from this starting point showed a few trends (not to be confused with research findings - this was not an official study). First, people are truly interested in systems that can save them time, do menial tasks for them, do practical tasks for them, and help manage the quantity of tasks they have to do. This informs the first of our guiding questions for this research:


How can robots make people’s lives EASIER?


The second trend of the form was that people are concerned about their home security/safety in general and also people with reservations about robots are generally concerned with the security of the robot system itself or becoming dependent on the robot. Responses like these inform the second and third questions:


How can robots make people’s lives SAFER?


How should robots INTERACT with humans?


These questions might seem extremely simple, almost obviously so, but the reason I have chosen them is because simple and clear guiding questions will allow the research to pull focus into the areas we really need to put time and effort into. For example, instead of focusing on developing an extremely robust software architecture, we would use existing frameworks such as ROS so the research can focus on the overall system design.


These questions were also developed from knowledge of the industry, by understanding previous startups (such as Kuri and Rethink Robotics), where they succeeded, and where they didn’t, a clearer picture of what people are looking for in a robotics system was developed.


Taking these guiding questions, I’ve spent a lot of my free time generating ideas and designs about what a future robot could look like. I spent time building models of my ideas and mapping out  how they could potentially function. The models and designs aren’t included in this abstract as they were mostly design exercises. They were not true prototypes based on actual user findings or extensively developed concepts of what a future robot can look like, but the basic framework from some of the designs could drive further research. 


wrap-up.

To summarize the above proposal, the future of robotics in the home will lie in the system-level design of a robot that implements existing technology (hardware and software) in a platform that makes living easier and safer in the modern world with an interface that makes it extremely easy to use and understand. 


references.

Scientific American - Bill Gates on the Future of Robotics “A Robot in Every Home”

http://agents.sci.brooklyn.cuny.edu/cc30.03/papers/robotineveryhome.pdf


Report Created from Google Form Surveys

https://docs.google.com/document/d/1epx8oHNOtN-wUNDS4s8OGNI5BeIgq4Q3Ovacv9WdY2w/edit#

3 views0 comments

Recent Posts

See All
© copyright 2019 | aditya mehrotra