2021 was a challenging year for everyone-including the team at Hoaloha Robotics. The impact of COVID not only required us to reschedule our user trials again but also created component shortages and affected working together in our office. However, despite these challenges we still persevered, continuing to make progress. By mid-year, with everyone vaccinated, we were able to resume working fulltime together in our office, using masking and testing protocols as necessary.
Much of our development focused on a top-to-bottom iteration of our hardware and further improvements to our user interface and content. A continual process of parallel hardware and software iteration allows us to improve on all fronts as well as balance where best to implement and improve certain functions. History in the industry has often demonstrated that hardware locked down without considering the constraints or requirements of the software that runs on it (or vice-versa) results in too many compromises and ultimately failure. Fortunately, in today’s age of laser cutting, 3D printing, and online custom design ordering, while not quite as fast as software changes, we have often been able to design, build, and test hardware and iterate as needed, enabling us to progressively learn how to improve with every cycle and move closer to completion .
Meanwhile, it was interesting to watch related developments in the industry. For example, while Boston Dynamics continued to dazzle everyone with videos of their robots dancing and doing parkour, they found themselves changing hands once again as SoftBank passed them on to Hyundai. Meanwhile, the SoftBank was also reported to have halted production of “Pepper” and laid off a significant part of their staff working on that project.
Google/Alphabet treated us to a peek at their Everyday Robot project, spawned from their “moonshot” labs, featuring a mobile wheeled platform with a single arm that they are attempting to use machine-learning to train to do useful tasks. Amazon also unveiled their rumored home robot, “Astro”. However, in both cases, it hasn’t been quite certain yet when either will be made available. But based on what they have shown and communicated so far, neither appears to be direct competition for what we are developing, so we plan to remain in “stealth” mode for now.
However, our objectives at Hoaloha Robotics have not changed. We continue to be focused on delivering a “companion” to augment and empower the growing senior segment of our population. This is often misinterpreted or mislabeled as “eldercare”, which would imply that our goal is to replace the need for human interaction. First and foremost, we do NOT think it appropriate or possible for anyone to create a robot as a device that could adequately substitute for human “care”. Such a term also tends to often carry with it connotations of “ageism”, a negative perspective of seniors that regards them as a frail or lesser part of our society. While aging may bring challenges in certain aspects, people generally do not prefer to wish to be regarded as inferior.
Instead, we see our mission in line with how technology has always been most successful, by enhancing and extending human potential and capabilities. Consider, for example, that most of us that have mobile phones likely have more phone numbers and email addresses than we keep in our brains, but none of us consider that to make us a lesser individual. We use the technology to augment our cognitive capacity. Similarly, we may use email, texting, or other forms of social media to help keep in contact with others. In the past this might have required a letter, phone call, or in-person visit. It is not that we cannot do these former tasks, but find the newer forms more efficient. Likewise, successful business executives typically have great administrative assistants and pro-golfers, experienced caddies to accompany them, not because they cannot do their jobs, but because these supportive roles better enable to be most effective.
Surprisingly, it is this essential element, that I will call the “value proposition” that has typically been missing or misidentified by most attempts at creating consumer robots. Instead, it is often assumed that calling something a robot will be sufficient to sell a product. Unfortunately, the term “robot” often works against that because most of us imagine robots that can do what we do. However, that remains a fantasy. Such abilities as the Jetsons’ “Rosie” had requires not only dexterous manipulation and autonomous mobility but also an intelligence sufficient to interpret the world and operate in it.
But even if it were possible to do this today, there is another important aspect to defining a value proposition, and that is the ability to deliver it at a price that people can afford. This too is a significant challenge because a robot like “Rosie” would require far more components and technologies than the conventional devices popular today. PCs, mobile phones, and tablets typically do not require perceiving their environment in order to do their tasks and generally rely on us to transport them around. Finally, we don’t expect them to manipulate objects in our world. So, while it may be possible to build the mechatronics to do many of these things, the cost and software needed to make them fully autonomous are not presently within reach.
The only truly successful category of consumer robots so far have been floor cleaning bots, in large part because they offer an appealing, but single function at a price that consumers are willing to pay. Scaling up from there requires not only increased number of components and capabilities but also cost. Finding that right combination is harder than it might seem and tends to also require on much more advanced software. For example, while several companies have demonstrated the ability to create cars that can mechanically drive autonomously, the software necessary to do it reliably and safely enough without any human supervision continues to be elusive (not to mention at a high cost).
Such increased capabilities also bring yet another challenge; the user interface, i.e. what a human must do to get a device to do all these functions. While keyboards and mice helped enable PCs to become ubiquitous, and likewise, touchscreens for our phone, these generally don’t fit for a robot you expect to interact with. Buttons, dials, or other controls might work for appliances, but have their limitations as well. Voice is a more natural approach, but while speech recognition has improved over the past decade, it still hasn’t proven to be a slam dunk. Shouting out voice commands like a drill sergeant is not how we typically communicate. Instead, dialogue is more how we use language, but that requires more than just converting audio into words. For words to be useful, they must translate into a shared sense of meaning, and such meaning may also the social nuances of how words are spoken and how they are delivered back to the user. Dialogue also requires an understanding of context, which may include not only information included in the current conversation, but previous conversations. For example, our ability to use pronouns relies on our ability to retain and resolve what they refer to.
Considering all these challenges and requirements, it is not surprising that most attempts at consumer robots have failed. Robots in factories and warehouses have succeed in large part because they are programmed to do a specific task with reliable repetition (and often with a constant source of power) and that their value is sufficient enough to command a cost ranging from tens to hundreds of thousands. In these scenarios, the high cost of the hardware is more easily offset by benefit they offer by doing repetitious or dangerous tasks more efficiently or safely. Further, many of these robots are connected to a constant source of power rather than having to rely on a limited supply carried onboard. This is starting to change as battery technologies have improved, opening the door to more logistical (delivery) functions that can be adequately performed.
For home scenarios, this all becomes much harder, as again, a robot would typically require more power for all its sensors and processors than your laptop, tablet, or phone. While you could increase the number or size of batteries, that usually increases both cost and weight. You might not mind that your robot vacuum may have only battery capacity to vacuum your entire house and you don’t expect it to be ever-present at your beck-and-call. At its price, if it can do an adequate job of sweeping up the crumbs or dust bunnies, and returning to park on its charging dock, you are likely satisfied.
All of these factors account for why consumer robots have proven such a challenge. But there is yet another important factor. It is that robots are a technology that takes time to develop and address these requirements. iRobot CEO, Colin Angle remarked that development of iRobot’s lawn mowing robot took 12 years to develop, and even then they chose not to release it now. Any company hoping to deliver a consumer robot in 3 years or less is just fooling themselves (and their potential investors). All new and innovative technologies take not only time to develop, but typically encounter unexpected issues that are difficult to fully anticipate. Bill Gates recently related how he set out to eliminate polio, come up with self-sustaining sanitation, and cheaper, non-polluting power generation. Despite his best intentions, sufficient funding, and intelligent people to work on these objectives, all these projects hit unexpected obstacles, suggesting it isn’t simply raw smarts or money that leads to success. Similarly many attempts at consumer robots have failed despite being well-funded and staffed with brilliant people. It takes time to bake a cake and have it turn out delicious.
So with all of this, you might ask if we are remain optimistic about what we are doing. The answer is a definite yes. Why? Because we understand that this requires more than building the hardware, though that also requires a significant investment. For consumer robots to succeed they must deliver a clear value proposition–desirable benefits at a price that people can afford and a more advanced user experience that makes it easy and delightful to interact with. But knowing this isn’t enough. It also requires innovation, integration, and a positive iterative process–in a word perseverance. For that continues to be driven by the need and opportunity to empower our growing senior population.
Stay tuned.