The pioneer of vacuum cleaner iRobot’s performance in the Chinese market is temporarily unsatisfactory, but CEO Colin Angel has his own determined technical route and strategic plan: he firmly believes that only vision-based intelligence that can collaborate with people is the home service robot. Ideal form: In the future, the house or “home” will become an intelligent robot. iRobot will take a place at the core of the smart home ecosystem by virtue of its vision and environmental understanding and the ability to move autonomously.
At first glance, this seems to be another case of localization failure-in the field of sweeping robots, iRobot, which has the largest market share in the world, among Chinese consumers, the popularity even needs to be squeezed into the top three.
Looking at it again, this seems to be a competition between capital and the company’s development imagination-the valuation of iRobot, which was founded 31 years ago, on Wall Street is less than one-third of that of Stone Technology, which was founded 7 years ago, on the Shanghai Stock Exchange.
However, in the view of iRobot CEO Colin Angle, these and other troughs are just the scenery that will be experienced in the journey of climbing the highest peak of the intelligent service robot group.
This robotics expert is also the entrepreneur who developed a no-investment and no-financing MIT start-up team into a listed company that created and defined the “household cleaning robot” product category, and firmly believes that there is only a tacit understanding based on visual understanding. Collaborative intelligence is the ideal form of home service robots, and iRobot is fully prepared for this.
From the initial creation of space exploration robots inspired by ants that “can work without thinking like humans”, Colin led iRobot to develop more than 100 robots, covering more than 20 industries such as aerospace, deep sea, security, medical care, and toys. Products have different levels of intelligence and different shapes and sizes, but the purpose is always the same, that is, they must be practical and the value provided must be higher than their manufacturing (or sales) cost.
In the intelligent system built by iRobot, intelligence is hierarchical. Except for a small part of the topmost layer, all bottom-layer behaviors do not require planning, similar to the subconscious response of human beings. “By first focusing on building the bottom-level behaviors, and turning as much intelligence into behaviors as possible, the top-level planning can be made simple enough to really work,” Colin told Heart of the Machine. It is this different understanding of machine intelligence and the way of building intelligence from the bottom up that laid the foundation of iRobot.
Although the arrival of the intelligent era made Colin express the need to rethink the definition of “robots”—from “tools” and “helpers” to “partners”, he can understand the intentions and needs of users and build a certain emotional connection with humans. But you can feel that what he called “emotional connection” can still make people use smart robots without emotional burden, just like you and I now use smart phones with peace of mind 7×24 hours. “This is the service robot that an aging society needs in the future,” Colin said. “Too much time, money and energy have been wasted to make humanoid robots.”
From the original defense contractor to today’s small home appliance giant, Colin’s goal now is to lead iRobot from a successful robotics company to a company that can completely solve the concept of “understanders”, and thus become smart in the future. The core of the home occupies a place. He believes that the form of human living has not been considered from the perspective of intelligence. In his vision, the house itself will become an intelligent robot in the future. The various Internet-connected IoT devices indoors and outdoors are just like the “home” robot. components, and the home cleaning robot Roomba, which roams among them, will become an important part of the smart home ecosystem by virtue of visual understanding and autonomous movement capabilities, like white blood cells that clean up germs and garbage in the human body.
During the interview, Colin mentioned Collaborative Intelligence “collaborative intelligence” many times, which reminds me of Rethink Robotics founded by his mentor and business partner Rodney Brooks, which is based on Sawyer & Baxter. Robots created a new product category and technical field of “industrial collaborative robots”, but ceased operations and sold assets in 2018. In Colin’s words, “forgot to become a great company on the way to becoming a great innovator.” .
In the increasingly competitive home service robot market, can iRobot accurately gain insights into user needs, as when Roomba was used to define sweeping robots, taking into account technological innovation and business development, and leading the future of smart homes with better products and services?
iRobot CEO Colin Angle (Colin Angle) and the company’s representative product, the latest Roomba S9+ sweeping robot
Lidar is a dead end, so no time is wasted on “harmful” technology
A big reason for iRobot’s current bottleneck in the Chinese market is that the product uses a pure visual navigation solution and does not use any laser sensors. From the consumer experience, it is obvious that the initial drawing time of the iRobot sweeping robot will be slightly less than that of competing products using lidar.
Commercially available mid-to-high-end household sweeping/wiping robots mainly have three types of navigation technology, one is laser navigation, the other is visual navigation, and the other is laser + visual navigation. From the naming method alone, it can be roughly inferred that, at least at this stage, laser + visual navigation must combine the advantages of both laser navigation and visual navigation, and this is indeed the solution adopted by the current hot-selling sweeping robot products. .
Laser navigation and visual navigation are two different navigation technologies, and the industry is inconclusive on the two technical routes mentioned above. However, there is a good reference case for the battle between pure visual route and lidar-Tesla’s autonomous driving development plan. Musk has repeatedly said in public that Lidar has no future (doomed) and that Tesla has had many fatal accidents, and he has not eased his tone.
In the use of lidar, Colin and Musk’s heroes have seen the same. “We developed products with lidar 25 years ago and knew that this was a dead end…. Lidar may be useful in the short term, but to achieve real intelligence, the camera is the correct sensor,” Colin said : “This is a strategic choice.”
He is well aware of the controversy of this decision. But for Colin, lidar not only bypasses the problems that must be solved to achieve true intelligence, but also poses a hidden danger like a buffer overflow. In the future, it will only consume more time and energy to eliminate this hidden danger. Therefore, although it is simple to add lidar to Roomba, he will never invest in technology that he knows is “harmful”.
“Maybe Lidar will convince me that something is there, but if I don’t solve the problem of visual understanding, I won’t know what that thing is.” He emphasized: “Any way to reduce understanding of the environment is a kind of interference.”
Colin previously stated in an interview with foreign media that iRobot is the second largest company in the world with connected robot products after Tesla. In 2020, iRobot launched the iRobot Genius home intelligence platform, which can be seen as a further upgrade to the unified management of tens of millions of cleaning robots distributed around the world. What a Roomba “learns” in a home will be Common to every Roomba, this is a collective intelligence that continues to iterate and evolve.
The various home-related data collected and accumulated by iRobot over the years are undoubtedly a powerful asset and one of the reasons for Colin’s confidence. “We have the most intelligent robot in the world,” he said. Like Tesla, iRobot upgrades its software every few months to add new features to the product. But unlike Tesla, Roomba is not “fatal” without lidar, and iRobot can rest assured to follow the pure visual route.
Turn 90% of intelligence into behavior and build “collaborative intelligence” from the bottom up
In the field of robotics, the challenge of developing a human-level cognitive computing model is to make a model that is both accurate and universal. Unfortunately, these two goals have always contradicted each other, because to establish a more realistic model of rational behavior, one has to consider computing. cost.
The key to iRobot’s initial success is to follow the concept of “robots can work without thinking like humans”, that is, adopting the so-called Subsumption Architecture. Different from the mainstream view at the time, this architecture does not guide the robot’s behavior by constructing a complete digital representation of the external world, but by decomposing the behavior into a series of sub-behaviors, allowing the robot to make real-time responses to the surrounding environment. Interaction and response. These sub-behaviors are organized into a hierarchical structure, each layer can achieve a specific level of behavior ability, and the combination of lower-level behaviors can achieve higher-level behaviors. The first generation of Roomba was born from this, basically without any planning.
Colin believes that this bottom-up approach to build intelligence is the foundation of iRobot, and this approach can also be extended to build human-level intelligence. He explained: “Intelligence can be regarded as a combination of low-level behavior and planning-based high-level thinking. If I can turn 90% of my intelligence into behavior, then I only need to plan in the thin layer above. … How much can I do with low-level behaviors? The more I do at low levels, the easier the difficult part will be.”
The difficult part is to build “collaborative intelligence”, which is also the ultimate goal of Colin to build service robots. This kind of intelligence allows robots and humans to cooperate and cooperate tacitly, thereby establishing trust, so that people can rest assured that robots can do more and more complex things.
Nowadays, more and more robotics researchers believe that it is necessary for robots to think rationally about humans-not only treat people as obstacles or perfect game players, but we need robots to take humanity into consideration so that they can better Cooperating well with humans means achieving the so-called “value alignment.” If successful, we will have the tools that can greatly improve the quality of life.
Achieving behavioral compatibility requires robots to predict human behaviors and understand how these human behaviors affect their own behavior, while also enabling humans to predict the next behaviors the robot will take. In addressing these challenges, Colin said that iRobot has made some progress, but there is still a long way to go. “We are at the forefront of building the intelligence required for smart homes. This is very exciting… Every day we try to figure out what is preventing a closer partnership between consumers and robots, and how to build trust. , Let the robot know how to respect and sympathize with consumers’ home living habits, and how to cultivate similar emotional connections between consumers and other devices in the home.”
“In order to build a collaborative partnership, you need to understand what you are collaborating to maintain. To a certain extent, this is not a breakthrough in artificial intelligence, but a breakthrough in learning and understanding of the context. This goes back to why I don’t like the problem of lasers,” Colin said, “because I want to create the richest environment to build my intelligence, and lasers are not helpful for this.”
During his 32 years in the field of robotics, Colin has seen many different problems, some of which are definitely solved by hardware solutions, but for today’s sweeping robots, he believes that the main challenge comes from intelligence. , Rather than cleaning efficiency in the moving process. “I don’t care that iRobot’s sweeping robot took a few more minutes to draw the map. I hope the robot can really understand the household it cleans.”
Building a robot is the ultimate test of a system design. It is necessary to understand and define the problem very carefully, and choose the right combination of software and hardware. Hardware will increase costs, system fragility and unnecessary complexity, while software not only can be upgraded regularly, but it is almost always cheaper. Therefore, iRobot will continue to build Colin’s ideal machine intelligence around vision, which he believes is not only applicable to Roomba, but also to broader robots and larger smart homes.
The future “home” will be an invisible intelligent robot
On the path to achieve the ultimate goal, Colin is very careful not to fall into a local minimum. He said that iRobot is not only thinking about how to deal with the simplest answer to the current question, but also constructing the best answer for today’s problems that can be effectively extended to tomorrow.
In the next 20 years or more, he will lead iRobot to work on solving the concept of “understanding home.” His ambition is obvious, because it is about human life, and it is an issue that cannot be ignored by any company that takes AI as the first or smart technology as its core selling point.
When it comes to smart homes, many people don’t think deeply about the meaning of the word. Ask a person with more than 20 smart devices at home, is your home smart? Usually the other party will say that they are not smart. Colin pointed out that this is because consumers don’t actually know what intelligence is, but if you ask another question: “Do you want your home to be a clean and orderly, safe and environmentally friendly, comfortable and secure place?” I believe the answer is just Will become: “Yes, I am willing to pay for such a change.”
Colin believes that “home” will eventually become an invisible intelligent robot, or at least will function as a system. The home here includes residences, living rooms, indoor and outdoor activity spaces, etc. It is very intelligent, and its task is to take care and protect the people living in it.
“There may be Amazon or Google in the future smart home system, but iRobot or a company similar to iRobot is still needed to improve it.” Colin smiled and said: “I don’t need to be the center, but without us, human The family cannot be intelligent.”
Today, many countries and regions, including China, are rapidly entering an aging society. What kind of home service robots do we need in the future? For this problem, Colin believes that humanoid robots are not the answer.
“The ideal model of a domestic service robot is a model that can complete its designated tasks in the most reliable, safe and efficient way. Humanoid robots are too expensive and too risky,” Colin said. In the future, there may be many different robots that solve specific problems at home, and these robots can work together long before we have the ability to build humanoid robots. In addition to entertainment, such as serving tourists in Disneyland, it is difficult for him to find a scene where humanoid robots can come in handy.
“The way to solve most daily problems is not to build a humanoid robot, and then let it operate devices or instruments originally designed for humans.” Colin said bluntly: “Trying to make a humanoid robot has wasted a lot of time, money and money. energy.”
Colin said that many companies have made great robots, but couldn’t find a reason for doing so. The work of Boston Dynamics is amazing. “But these robots have no real value. If I were the CEO of Boston Dynamics, I would make a pet robot. I would use all these amazing dynamic control and jumping techniques to make one. Puppy, it can run over, jump on my lap, and I can touch it.” Colin said: “The most valuable application of Boston Dynamics’ technology can be to create robot pets to solve the problem of loneliness.”
He believes that Google is also a little too focused on innovation, spending too much money on the moon project. Colin emphasized that how to balance technological innovation and business operations is something that companies, especially technology start-ups, must always keep in mind.
Adhering to cutting-edge technological innovation based on pragmatism, iRobot has survived for 31 years and has become a name that cannot be ignored in the global robotics industry. When asked how he hopes to be remembered, Colin did not hesitate to say that he considers himself a Builder “builder”, and he hopes to be remembered because he has produced practical robotics technology and established a practical robotics industry. .
In the Chinese context, Builder feels more like “filling holes”, building things to solve problems. Colin said that iRobot hopes to become a company that builds collaborative intelligence and creates robots that can help humans live independently for a longer time. For him, this will be a lifelong journey. In this sense, the company’s latest slogan really conveyed his aspirations: iRobot, so you can human.
The Links: FLC48SXC8V-05 QM150DY-HK INFIGBT
0 Comments for “iRobot CEO Colin Angel: Lidar is a dead end, and the future home is an intelligent robot”