For many years, we have believed that robots will be as pervasive as smartphones throughout human society.
As early as February 2007, Bill Gates predicted in the "Global Science" magazine that in the near future, every family will have a robot.
But the reality is that today, this dream is still far from reality.
Although the research and development, manufacturing and application of robots have always been highly valued by the government and capital, China has been the world's largest robot market for eight consecutive years since 2013.
However, unlike industrial robots that mainly work in "unmanned environments" and pay more attention to efficiency and accuracy, the application scenarios of service robots require close contact with people, so they pay more attention to the "flexibility" and safety of the service process.
Therefore, service robots need to be lighter, smarter, and more capable of environmental perception and human-computer interaction.
Service robots must rely on comprehensive breakthroughs in different fields of technology such as new materials, artificial intelligence, flexible joints, sensors, etc., in order to truly achieve product maturity.
At the same time, more complex technical requirements mean higher costs. This further restricts the commercial popularity of service robots.
【2】
However, the situation has gradually changed.
On the demand side, since 2019, the raging of the new crown epidemic has made "contactless service" a rigid demand. From hospitals to hotels, from distribution to inspections, in the process of epidemic prevention and control and resumption of work and production, a large number of Service robots take to work.
Even in recent years, there have been more and more representatives of the industry, suggesting that robots should be used to replace manual repetitive labor under the conditions of meeting the requirements of economy, technology and reliability, so as to alleviate the increasing "labor shortage".
More importantly, on the supply side, there are solutions for the two major bottlenecks in the maturity of service robots.
The first bottleneck is the level of intelligence.
In the past, due to the limitation of artificial intelligence capabilities, service robots could only be like feature phones, engaged in limited fixed jobs, food delivery people could not clean, and inspections could not cut grapes.
Now, with the continuous breakthrough of new materials, high-precision control and software algorithms, artificial intelligence has made rapid progress in the fields of deep learning, knowledge graph, scene recognition, human-computer interaction, reasoning and cognition, and service robots are about to achieve a subversive upgrade. :
Just like smartphones replacing feature phones, the currently widely used service robots with a single function will gradually be replaced by service robots with general intelligence that can provide multi-functional, multi-modal service experience.
The upgrade of China's communication infrastructure has made the landing of "cloud intelligent robots" possible.
The movement of the robot needs to rely on joints. In the past, with the increase of joints, the amount of computation would continue to increase, which would consume a large amount of computing power and energy of the robot, resulting in an increase in the size, energy consumption and cost of the robot, which greatly affected the application and popularization of service robots.
The "cloud intelligent robot" is the architecture proposed by the Chinese robot unicorn company Datsun. It separates the "brain" of the robot responsible for computing work from the ontology, and puts the "brain" in the cloud, and the robot ontology is mainly responsible for perception and service.
In this way, the "upper limit" of traditional service robots is completely broken: only one communication network is needed to connect each joint, and the entire system becomes an infinitely scalable system that can make any complex robot.
The premise of this idea is to establish a complete and high-speed "cloud-network-terminal" system. Through a powerful cloud brain, a high-speed and reliable neural network, and a large number of robot bodies, the whole system can be completed for us. Serve.
A major development of the 5G network is to shorten the transmission delay of network signals to less than 10 milliseconds-correspondingly, the signal transmission speed of human nerves, the delay from the brain to the joints and muscles generally needs 180 ~240ms.
At the same time, the "East Data and Western Calculation" project officially launched by the state this year has also put forward very high technical requirements for the data transmission of the backbone network. In the Tencent data center, the delay does not exceed 20 milliseconds in principle.
This means that, through the "brain" a thousand kilometers away, to support the intelligent robot to serve us, its response speed can still be faster than the neural response of our human body.
The second bottleneck is the flexible service capability of robots.
In the past, the vast majority of service robots had feet but no hands, with rollers to move and even interact, but most of them did not have flexible arms and could not serve us like real people.
However, facing the future, no matter whether it is delivering goods, taking care of the elderly, organizing housework, or handling farm work, the more intelligent the robot, the more it needs to have a strong "hands-on ability".
This bottleneck is the last barrier to overcoming high-end robots: in order for service robots to obtain flexible arms and realize complex combined actions, they must have intelligent, miniaturized, and high-strength flexible joints.
This core component accounts for 70% of the cost of intelligent robots. It is also a shortcoming of my country's robot industry. It has long been monopolized by the "four families" of foreign robots.
And now, domestic robot companies have also made up for this key shortcoming through independent research and development.
In 2009, in the Maqiao Artificial Intelligence Innovation Experimental Zone in Minhang, Shanghai, an intelligent robot base covering an area of 243 acres has been erected.
One of the core products mass-produced by this base is the SCA (Smart Flexible Joint) independently developed and produced by Dash. Under the same performance, its volume is only 1/10 of the traditional servo system, and the minimum diameter can be 35 mm, which can make the robot smaller and more flexible in movement.
Relying on the SCA, Dash was approved as a strategic emerging industry project in Shanghai in 2021, which is the project with the most support since the establishment of the special project. At present, more than 100 ecological customers are using Cloud SCA, including medical robots, rehabilitation robots, special robots, etc.
Previously, Dash has successfully developed and manufactured Ginger, a humanoid service robot with 34 degrees of freedom in 2019. It has a pair of flexible hands and can perform various actions.
"The human skeleton has about 350 degrees of freedom. In the future, as long as there is a network with a high enough bandwidth, a fast enough delay, and an IP address for each joint, we can completely create a dual A cloud-based intelligent robot in the shape of a human being, Dada aims to create a home nanny robot for every family." said Huang Xiaoqing, founder of Dada.
【3】
It is worth noting that in the wave of development of service robots, Chinese enterprises are expected to occupy a more important industrial position.
Prior to this, due to the late start, lack of technology accumulation and weak industrial foundation, China's industrial robot industry had to rely on foreign companies for a large number of core technologies and core components, and high-end supply was lacking. Until now, it is still necessary to keep catching up.
But in the field of service robots, China has come to the same starting line with foreign companies, and even established a leading edge in some fields such as cloud, network, and intelligence.
One of the most representative examples is that as early as 2015, Dada first proposed the concept of "cloud robot" and fully deployed it. At present, Dash has begun commercialization and has a first-mover advantage in multiple vertical industries such as commercial retail, public health, communications, education, public safety, power and energy, agriculture, and sanitation, and has huge room for growth.
Now, Dachshund has successfully built the robot's cloud brain operating system "Hairui" and the robot virtual development platform "Haiyuan World" - the metaverse of robots.
The essence of the Hai Rui system is exactly the "cloud-network-terminal" architecture built by Dash for cloud service robots. It has established a brain with general intelligence capabilities, a safe and efficient neural network, and has been successfully verified. The cloud intelligent robot system.
Haiyuan World is the working and learning environment for robots to understand the physical world, intelligent training and virtual-real interaction. With the support of cloud computing power, millions of digital twin robots continue to perform deep reinforcement learning in large-scale simulation scenarios in the Haiyuan world, gradually forming the intelligence of cloud robots. Then through the combination of virtual and real technology, the robot can work in the real world.
At the same time, Dash's Hairui operating system also provides developers with an integrated application development kit RDK for robot application development, namely HARIX RDK. It provides a complete visual development tool. Developers can realize low-code application development by dragging and dropping directly. In this way, developers can completely get rid of physical robots and develop various robot applications quickly and at low cost, which greatly reduces the number of robots. threshold for application development.
"Even, people who don't know how to write programs can do the application development of robots." Huang Xiaoqing said.
The purpose of launching the Hai Rui system is to open the cloud intelligent robot architecture to the entire industry.
In the past, if an entrepreneurial team wanted to develop an intelligent robot, it needed to first consider what kind of chips, sensors, joints, servers, radars, and image modules to use, how they would be assembled, what product form they would be, and then never again. To develop drivers and learning algorithms, and then to verify and adjust and optimize step by step...not only time-consuming and laborious, but also costly and risky.
Now, relying on the Hai Rui operating system, third-party robot companies can quickly complete the simulation modeling, application development and virtual verification of the robot on the virtual platform by simply accessing the cloud brain of Dash. Just make a real robot.
"Our positioning is to be the architect of the ecology." Huang Xiaoqing said that the entire intelligent robot industry is an extremely long and complex industrial chain. Among them, most of the fields that other people can do will not be involved in Dakar. "For example, We also looked for smart flexible joints first, but no one in China could make them, so we had to do it ourselves to fill this gap for the ecology.”
Now, Dash has cooperated with many industry partners and universities, serving more than 2,400 customers in the industry chain. "Datalink will continue to promote the construction of the cloud robot industry ecosystem and the large-scale development of the industry." Huang Xiaoqing said.
【4】
At present, a common perception in the industry is that the next ten years will surely become the "golden decade" of service robots.
If the epidemic has opened a window for the application of service robots, then the severe challenges of an aging population and a serious shortage of labor, as major challenges for Chinese society in the future, will surely promote more industrial robots and service robots, more and more Intelligence, which is becoming more and more pervasive, helps us do more work.
In this process, with the introduction of information and communication, biotechnology, new energy, new materials, artificial intelligence and other technologies, the robot industry will also stand on the cusp of taking off, ushering in a qualitative change from function to intelligence, continuous upgrading and leapfrog development , and finally achieve large-scale application popularization in the whole society.
Perhaps, in the near future, just like Bill Gates' prediction, we will usher in the era of "robots in every house".
At that time, the cloud intelligent robot will become the third computer of human beings after PC and mobile phone. "Huang Xiaoqing said.
