At the 2021 World Robot Conference exhibition hall, a reporter saw a collaborative robot demonstrating the strawberry picking process at its booth. It was explained that similar collaborative robots are already being used in orchards in New Zealand to help fruit farmers harvest kiwifruit.
So, compared to the relatively fixed hardness, size, and position of the strawberry models in the exhibition hall, how does the robot determine which kiwis, varying in shape, size, and ripeness, can be picked in a real orchard setting, and which cannot? Which need to be handled gently, and which require "force" to be picked?
According to the staff at the booth, this is because the collaborative robots are equipped with industrial cameras to determine the size, shape, and ripeness of the fruit, and the harvesting is automated through algorithm optimization.
In fact, with the iteration of robotics technology, robots are no longer just "giant machines" in factories, but are gradually becoming "flexible assistants" in people's lives. The exhibition hall also featured many similar robots, operating with either one arm or two arms in coordination, smoothly performing tasks such as grabbing eggs, sorting packages, delivering water bottles, creating latte art, and writing the character "福" (fortune). The reporter learned that these robots' current "primary jobs" are mostly in the production and manufacturing of automobiles, 3C products, and other fields. They are all classified as collaborative robots, evolved from traditional industrial robots.
What are collaborative robots? Yang Shuai, vice president of Hunan Ruisenke Robotics Technology Co., Ltd., pointed out in an interview with CNR.cn that "at the academic research level, robots that can collaborate with people and interact with the environment are called co-existing robots, while in industrial applications they are called collaborative robots. In fact, both emphasize the robot's interactive capabilities."
Sun Fuchun, a professor in the Department of Computer Science and Technology at Tsinghua University, also told CNR.cn, "The integration of humans and machines is like collaboration between two people. If the two people have different ideas and approaches, it will be difficult for them to cooperate. The collaboration between machines and humans requires coordination between humans and machines in terms of perception and communication, which requires AI algorithms to accomplish."
"For example, people who wear prostheses want them to truly become an indispensable part of their bodies and be able to feel the touch of the prosthesis. This can be achieved through a neural interface. That is, by connecting the human nerves and the electromechanical signals of the prosthesis, after neural decoding, the brain's instructions are transmitted to the mechanical prosthesis, and after the prosthesis senses the touch, it also transmits the touch signals back to the brain through the neural interface," Sun Fuchun told reporters. Another method is cross-modal triggering, which allows the robot to learn features such as the shape and touch of a large number of objects through cross-feature learning and establish connections in its "brain".
"For example, when we smell the aroma of braised pork, we think of braised pork because we have eaten it before and have a basic understanding of its shape and taste," Sun Fuchun said. He added that robots also need to learn through algorithms to become "smarter."
It is understood that Professor Sun Fuchun's team showcased several advanced technologies at the exhibition, including a throat swab collection robot, a five-finger dexterous hand, a ward patrol robot, and a mobile dual-arm robot. Among them, the autonomous throat swab collection robot system employs force feedback and compliant control strategies to ensure the safety of the collection process. It combines deep neural networks to design a pharyngeal region recognition algorithm and a motion path autonomous planning algorithm, making collection more accurate. The entire collection time is less than one minute per person, completing the workload of 2-3 nurses, with high sampling efficiency and an effectiveness rate exceeding 94%. Furthermore, the system features automatic throat swab loading and unloading, intelligent voice prompts, and system error alarms for enhanced ease of use. A newly designed mouthpiece with human-like wiping motions ensures the reliability of the collection process. It is reported that the system has already undergone practical testing in multiple locations, including the Suzhou Fangcang Hospital and the Nantong CDC.
Why are collaborative robots considered an iterative upgrade of industrial robots? Yang Shuai told reporters that traditional industrial robots are often installed inside protective enclosures, preventing direct contact with humans and requiring specialized personnel for programming, making them inaccessible to the average person. Therefore, for robots to "adapt" to human environments and better assist humans, they need autonomous environmental perception and real-time interaction capabilities. It is against this backdrop that collaborative robots have emerged.
Taking Risen's Baxter dual-arm collaborative robot, showcased at this conference, as an example, it utilizes flexible joint technology and incorporates multiple sensors, including vision and force sensors, enabling it to safely interact with people and the environment. Furthermore, by employing a traction-based teaching method accessible to ordinary users, the robot can be trained to perform a task within minutes.
Yang Shuai stated that collaborative robots are currently mainly used in the industrial sector to compensate for the shortcomings of traditional industrial robots. However, in recent years, more and more collaborative robots have begun to be applied to fields outside of manufacturing, including healthcare, logistics, security, and retail. "Collaborative robots have greatly expanded the application capabilities and boundaries of industrial robots. The robots of the future will definitely all be robots capable of human-machine collaboration and safe interaction. This is an inevitable trend."
According to the "2020 Blue Book on the Development of the Collaborative Robot Industry," collaborative robots accounted for 5.36% of China's total sales in 2019, an increase of 1.32 percentage points year-on-year. It is estimated that the sales share of collaborative robots in China will reach more than 6.5% in 2020 and more than 7.5% in 2021.
Wang Yaonan, an academician of the Chinese Academy of Engineering, gave a simple definition of a robot: "Whether it is an industrial robot, a service robot or a special robot, simply put, a robot is an integrated mechanical device. In recent years, the functions of perception, dialogue, decision-making and control have been added to robots, making them more intelligent."
Meanwhile, Wang Yaonan pointed out regarding the future trends of robots that the service robots most in demand in academia, industry, and manufacturing will increasingly evolve into Collaborative Robots 3.0. This requires robots to possess cognitive learning, human-computer interaction, semantic analysis, and especially natural language understanding, as humans must understand machine language to collaborate with them. "The development of robots is a continuous process of iteration. Through technological innovation, robots will become smarter, more efficient, more reliable, and provide better services to humanity."
