background image

Amazon, Google, and Robots! Oh My!

By now you have likely read about the interesting news about robotics developments related to Amazon and Google. This has raised several inquiries about the significance and its impact on what Hoaloha Robotics is doing.

First, with regards to Amazon’s suggestion that they might be delivering packages by autonomous drones within the next 5 years, the concept of flying drones delivering packages to your doorstep seems fraught with challenges that make this highly unlikely. However, I take it somewhat as more of a publicity move to help stimulate online purchasing for the holidays. In addition, I also think it enables Amazon to highlight the innovations they are thinking about to make deliveries to you even faster. What was surprising is that the video showed the drone flying out of what looked like a conventional warehouse. It might have been more impressive if they had shown it in a Kiva Systems-equipped warehouse, with a Rethink Robotics Baxter robot doing the picking and packaging. Considering that Amazon already owns Kiva Systems and that Bezos has had a private investment in Rethink, this latter scenario seems much less far-fetched. So I have no doubt Amazon will marshal robots to make delivery to your door faster in the future. I am just not yet convinced it will be via flying drone.

Google’s announcements were more interesting. First, it clarified what the next role for Andy Rubin (who had stepped down from leading Android development)  will be and that the company believes that the timing is right to aggressively invest robotics as they acquired a number of robotics technology companies. Some were spin-outs of Willow Garage, co-founded by Scott Hassan (a former architect at Google and now CEO of Suitable Technologies), though the most recent and notable of was Boston Dynamics.

If you are not familiar with Boston Dynamics, this company has developed a number of large legged robots including Big Dog, a robot that looks somewhat like a headless mechanical horse or deer. Designed primarily as a possible future transport of equipment for military troops, its agility in being able to walk across a variety of surfaces, including icy pavement or rugged forest terrains, is impressive. It has been followed by other equally innovative robots like the Cheetah (named for its ability to run like one) and the Atlas, a humanoid robot that are being used by some teams competing in the latest DARPA Robotics Challenge.

While the portfolio of acquisitions demonstrates that Rubin and Google are serious about robotics, it remains a question as to what they plan to create with all these technologies. It seems unlikely the goal will be to build any consumer/personal robots, at least for the near term. So iRobot’s dominance in the vacuum cleaner robot space is probably safe. All the acquired companies develop relatively high-end robotics technologies, i.e. what they produce are not cheap.

There is also some commonality in that many of these companies have been oriented toward mobility and manipulation.  It is therefore logical to assume their plans may to build something with such characteristics. Perhaps a part of this may have been motivated by Rethink Robotics’ early success with its Baxter robot and the growing number of companies building lighter weight industrial robot manipulators that can also be safely used around people.

That said, commercial robots that are both autonomously mobile and also support truly adaptive manipulation (able adapt to a number of tasks and environments) is really not here yet. While  industrial robots have been doing manipulation for years, there are doing very repetitive work from stationary positions or within a well defined, limited space. Similarly, as impressive as Baxter is, it currently operated from only a stationary platform and it has primarily designed simply to replicate the same pattern, similar to how its factory cousins do. Combining the idea of doing a variety of adaptive manipulation tasks with moving within in the dynamics of a human environment is still a hot topic being pursued in university research labs. So while you may see online videos of robots folding towels, sorting socks, or serving drinks, these demonstrations are typically highly scripted.

While engineering improvements will accelerate, I don’t believe we are likely to see  affordable consumer applications of this level of technology much earlier than we see self-driving cars that, while has been demonstrated, likely has at least a 10 year runway yet to go. Don’t get me wrong, autonomously mobile, adaptive manipulation is a coming technology and when it arrives it will be a game changer, but it will take time to get the costs down and the software required to get there, not to mention working through the other related issues.

How might these announcements affect Hoaloha? While these companies represented here are developing an incredibly powerful set of technologies that could be used to support a wide variety of applications in the future, nothing here appears to be directed in a way that might compete with Hoaloha’s objectives and its development efforts in the near term. You can see how many of the technologies Google acquired are being tested in the DARPA Robotics Challenge; again because of the costs to build robots like this, these robots are being applied still to the scenarios where a robot might replace a human in a dangerous or harmful environment. So the objectives are still toward the 3 D’s (dull, dirty, and dangerous) tasks that have classical defined the role of robots so far. 

I also see little in Rubin’s acquisitions in terms of human-robotic interaction (HRI). While some of the technologies they acquired can be used to create humanoid robots, the technologies are more oriented to extend or replace human capability rather than to interact with people, and they likely will still require a human operator to control or monitor them in the short-term. Thus the likely target scenarios are  more like how John Deere’s autonomous harvester combine, already on the market for a number of years, but it still requires a human operator to be present to monitor its operations in the same way as pilots are required to monitor a jet’s auto-pilot. Some might argue that the inclusion in the acquisition of Aaron Edsinger’s Meka Robotics does give Google some resources here. While it is true that Meka’s robot “face” has been used to create expressive robots, designing a humanoid looking robot is not the same as designing human interaction. A wonderfully articulated puppet does not provide its “magic” without a skillful puppeteer that controls it. The mechatronics of a robot may be an important enabler, but requires more to provide the experience.

Further, Hoaloha’s design starts a little differently. We start with our focus on creating a core value proposition for our target audience, developing autonomous robot companions to assist and empower the growing population who face challenges from the effects of aging, disability, and chronic disease. To that we add the priority of creating a user interface that is both natural and easy to use.  This drives much of the decisions we make on technology and hardware, rather than the other way around. We see our success dependent more on what value we deliver and the quality of the user experience than on how amazing our technology can be.

However, these announcements do have a positive effect on what Hoaloha is doing. Amazon and Google are well known companies that have been successful innovators of technology. So their investments highlight that the age of robotics is near and an increasing reality. As a rising tide floats all boats, so too this helps pave the way for what we are doing.