Interview conducted by Paul Rogers for Drive the District
Q. First-off Dr. Earon, can you please share a little bit about your background in robotics.
A. My background in robotics has focused, very heavily, on giving remote systems the onboard intelligence to effectively do the things that machines are good at, like fast calculation, data collection, analysis, and mobility and navigation. This makes the robots much easier to use and allows humans to do the things that we're good at such as high level decision making, determining what are valuable questions that the machines can answer and understanding how this information fits into our broader understanding of what is going on in the world.
I have been working in UAVs for almost 10 years now. My current focus, at PrecisionHawk, is on a easy to use system that will go out and get the best possible data without the need for human involvement in control, flight, etc. A user just needs to throw the plane, and it won't come down unless it gets good data or decides that the conditions aren't sufficient to get that data efficiently.
Prior to this, I worked on fleets of cooperating vehicles that could communicate in the air and share the task of search and rescue. For example, with each different type of vehicle being able to use its unique abilities (hovering, driving, fast flight) for the mission. I also designed novel vehicle types and control for them.
My background, and the focus of my doctoral work, was in space robotics for planetary exploration. The focus there being to give the robots artificial intelligence to effectively, and safely, navigate on another planet and perform useful science for us. Driving a mobile robot on Mars by hand from Earth, while possible, is not the most effective way to collect information.
Q. So, I understand that there are currently close to one million robots at work around the world. What sorts of roles are these machines performing and in what sorts of contexts?
A. Robots are finding their way into a great many applications. The most common use is still probably industrial robotics and manipulators used in manufacturing processes and dangerous work environments such as nuclear reactors. Robotic manipulators have been increasingly used for simulation and training, surgery, bomb disposal, and science and exploration on other planets, though perhaps the latter don't count towards the "global" total once they're launched.
Ground vehicles are showing all over the place as well. They have been used extensively for militaries and police forces all over the world and have been a great asset to keeping people away from very dangerous situations. Some radio control vehicles now have very sophisticated systems onboard. It's now possible to buy a robotic vacuum the size of a frisbee that will drive around your house sweeping the floor and returning to its charger when needed.
There are remote sensor buoys in the oceans and submersible "gliders" that can patrol the oceans for months collecting a huge variety of key information on everything from carbon absorption to salt content and temperature.
In certain parts of the world demographics have led to a real concern in how to deal with people who need assistance with day-to-day living, and there is a great deal of development in that front. Countries like Japan have been working very hard to create systems that are not only very effective, but designed to be accepted by the person they are assisting. That means being able to understand, and even respond with, relevant emotional responses, visual and non-verbal cues such as "smiling" or furrowing a brow.
There have been unmanned aircraft since World War 2 with more widespread use starting in the 80's. Since the late 90's that has continued to grow and now we're in the middle of an explosion of airborne systems with very capable aircraft available to anyone. A few hundred dollars and get you a hovering UAV that can stream video back to your smart phone and let you zap imaginary monsters in the video feed. The use of aerial systems for agriculture is poised to explode. Large parts of the world are already using these systems to great benefit over crops, livestock, etc. This is in large part what PrecisionHawk has been built upon. However, every day more and more applications are being brought to us, and we’ve expanded our technology and capabilities to reach those needs.
Robotics is probably making its greatest advances, and adoption, in places that we don't expect. It is not unreasonable to say that a fridge that can monitor its contents and order new groceries, or a thermostat or household lighting/heating/security system that can learn and adapt to our daily behavior is a robotic system. Our phones have the computation power to monitor a camera imagine in real-time and locate and identify faces, and then snap a picture as soon as the face is smiling.
Q. How are robots proliferating right now and in which areas (industrial? medical? domestic?)
A. Robotics is proliferating very quickly and across the application spectrum. Certainly industrial systems are getting far more effective and capable with very precise control and force feedback allowing their use in more applications, and do them better. The same is true in the medical fields, certainly, and for the same reasons. I think that the greatest growth, though, will be in the civilian spaces. This includes the use of robotic vacuums, thermostats, but the greatest economic potential, in the near term, is in the, as yet, largely untapped markets for rapid, on-demand, remote sensing. Agriculture is a great example of this. It is an industry that is very data intensive, but has specific needs as to timeliness of that data, ease of use, robustness and speed. Farms don't need planes. They need an end-to-end method of collecting, processing, and distributing a variety of information. Intelligent robotics is very well suited to be the start of that information pipeline.
Many people have thought of robotics as being dedicated to "dull, dirty, dangerous" jobs. It is certainly something that much of the industry had been spouting for several years. I think that is false. To limit these systems to those roles does a disservice to the potential of intelligent platforms in our daily lives. Robotics need to be able to make an economic case for themselves and carve out niches in markets where they are truly the best solution to problems. There are certainly a great many problems in the world that can be efficiently and effectively solved by robots. Again, agriculture is a great example. Fast, easy, on-demand spectral data collected daily from a soy field is a huge benefit to growers, researchers and analysts. It is a great way for producers to understand how best to operate in a world where we must constantly be doing more with less.
Certainly, bomb disposal is a great use for robotics, but they have long ago moved on into new areas.
Q. What are the technological breakthroughs that are facilitating the proliferation of robots? And what are the technological challenges that are limiting the spread of robotics?
A. There are several key technologies to robotics that are crucial to their effectiveness. They include sensors, power, computation, and materials. We're all familiar with how rapidly computers are evolving. With each new release they are smaller, better, faster, and cheaper. All of these, and the technologies around them, are advancing incredibly quickly today. A battery that was state of the art last year, is woefully inadequate compared to new capabilities. Look at cell phone batteries and the "arms race" to keep up with the processors they're powering.
The real benefit is that all of these are advancing now under their momentum. Robotics researchers and developers don't need to spend all their efforts creating a new battery or sensor technology, they can focus on the intelligence of the system, which is what will really allow them to do amazing things. For these systems to be effective, they need to be smart enough to adapt to the world and change their behavior in response to whatever conditions arise. These advances in sensors, materials, etc, allow robotics developers to spend far more time on that intelligence and how to solve real-world problems.
One key requirement for robotic systems is robustness. When robots have made great inroads is when they have been solid, reliable, and dependable, day-in and day-out. For robotics to be extremely widespread, they need to be very robust. Many of these platforms are very sophisticated and complicated. We need to give these systems the capability to be as reliable as the other components in our day-to-day lives. Look at cars for example. The fact that we can drive a car for several years in all sorts of conditions with only minimal upkeep, is astounding. That's where robotics needs to get to. Imagine a robotic lawnmower that was as robust and effective as a goat.
Q. Google recently snapped-up 8 robotics companies in just 6 months? Does this surprise you? And what do you think their thinking is behind this aggressive acquisition of such companies?
A. The consolidation of the industry is not a surprise, though the pace at which Google does things is often quite startling. The robotics and unmanned aircraft industries have many parallels to manned aviation in the 1920s. There was a spark of innovation, a flurry of interest, then the economic value and business cases started being made. The systems got more capable and robust, and more and more applications were found where these systems were able to shine. All the while, there was a consolidation in the players in the space as smaller producers, developers, and researchers where increasingly absorbed or replaced with larger organizations that were able to put serious resources behind it. We went from flying 120 feet in 1903 to landing a human on the moon 66 years later. When you imagine that kind of potential advance with today's rate of technology improvement, it's easy to see the value. Particularly as new applications and value propositions are being made every day.
Q. Are we really going to having our Christmas gifts dropped-off my unmanned aircraft?
A. I think Santa may have an issue with all our Christmas presents being delivered by a UAV, but wholesale delivery is a bit far off yet. I don’t want to dismiss the fact that Amazon is looking into the concept…Certain aspects of their business model really benefit from these capabilities. I think a great many companies, and some that might be unexpected, are looking to these new technologies because they can be very disruptive. Once these systems have the intelligence and capabilities, the applications are really open.
The other interesting, and very valuable, aspect of the Amazon headlines is that it really put the issue of widespread use and adoption at the forefront of people’s conversations. It really opens up the discussion and has people saying, "You know what, maybe that is possible."
Q. What other specific cutting-edge developments are there in the world of robotics right now, and who is leading the way? For example, I know that Carnegie Mellon University’s “HERB” robotic butler has made some waves.
A. I think as robotics gets becomes more specialized, there are cutting edge developments in each specific field that allow for the best job to be accomplished in the field at hand. PrecisionHawk's solutions for agriculture are incredibly sophisticated and specialized for supporting agricultural users. Aerovironment has developed a hummingbird sized (and styled) UAV for DARPA. In terms of legged locomotion, Boston Dynamics has been developing incredibly advanced systems. Developers in Japan have been leading the development of humanoid robots for years, and the venerable ASIMO and later developments have continued to turn heads. Honestly, one of the things that astound me the most, is seeing these technologies show up on toy shelves for less than $100. For example, a remotely piloted helicopter, three inches long that has an infrared emitter to "tag" another helicopter for $20… Amazing.
Q. What is your vision for the proliferation of robots over, say, the next 10 years and what will be behind these? Where are robots going to be used more; and which new area of life will robots “invade”?
A. I think that robotics will continue to proliferate, and with increasing speed, through our daily lives. I think that the biggest change that we will start to see is institutions and infrastructure being adapted to better accommodate robotics and take better advantage of them. Roads that are designed for driverless cars are not too far off. Improved mobile data solutions so that remote systems can be transferring huge amounts of data in real time, no matter where they are, will become critical. Sensors and what they sense will co-evolve more. Imagine if we start to see plants that are bred to exhibit specific signals in different light spectra, which happen to matchup with the spectra that aerial vehicles are readily able to carry.
I believe that the drive will come from people starting to think outside the box in their fields as to what can be done. At PrecisionHawk, we have users come to us all the time with new ideas and new applications that we never even imagined.
That, and our alarm clocks will start to get better at learning how to best wake us up.
Q. Finally, when will our average reader likely be directly impacted by (and directly interact with) robots and in what ways? For example, when will a robot be grabbing us a beer from the fridge; driving our car or walking our dog; or acting as anesthesiologist next time we have surgery?
A. I think that the average person is already being impacted by, and interacting with, robotic systems. Our phones, our cameras, even our washing machines have more intelligence and computational power than the systems that put people on the moon. Thermostats that self regulate based on how we live, Google glasses, watches that talk to our phones are just the tip of the iceberg in terms of how we will be interacting with these systems.
Ford has taken a production car and outfitted it so that will automatically brake in response to an obstacle. If it won't be able to stop in time, it will steer around the obstacle. All of this will happen without input from the human driver.
Most of these are available now, and some are just around the corner.
If you want to think of a more popular, traditional view of a “robot” as something that moves and can move things around, I think that people will get introduced to these systems when they are needed, or make sense to use. Will they walk to the grocery store to pick up dinner for us? Not soon, but will your robot vacuum cleaner be able to able to email you that it got stuck eating one of your socks? We just need to wait for the app.