Ken Goldberg, professor of robotics at Berkeley, Calif., looked at his face. He kept rubbing his coffee cup and said in his mouth: “How difficult is it for robots to master such data.†Nowadays, artificial intelligence can handle complicated perceptions with ease. Work such as assisting legal and medical research, but for robots, picking up the clothes that fall on the ground is still a fantasy. Berkeley, Cornell and other universities and Amazon, Toyota and other companies are constantly working hard to make the manipulators have the same agility as manpower. If someone succeeds one day, it will undoubtedly trigger a new round of robotic revolution. These smart machines will further release social productivity. 
It is true that machines have been in our lives for centuries, but the work they can do is rather limited. "They are all placed in fixed locations and do all kinds of mechanical tasks repeatedly," Goldberg said. However, once out of the factory, the machine becomes helpless in unstructured environments such as chaotic rooms and busy warehouses. Grab the trick “Taking an object sounds simple and humans can do it easily without even thinking, but for robots, this action is subtle and elusive,†says Goldberg. But if you carefully consider this process, it really depends on a very complex network in our brains. Take the mug, for example, the human brain will automatically calculate how to hold the cup the most stable, and it will be clearly defined where each finger should be placed. Through human evolution, the brain has its own highly customized processing conventions. "Although I had never seen the pen on the table, I knew I could easily pick it up," Goldberg said. “In the process of taking the pen, the brain regained a similar experience and passed it back to both hands.†Now, Goldberg is learning this trick with his students. To this end, they have specially built a network database called Dexterity Network (Agile Network), where about 10,000 3D virtual items are stored. The scale of database storage of virtual items may gradually expand to a million. When I visited Goldberg's laboratory in September last year, he put a lot of grotesque 3D printing models in front of me. Goldberg asked me to try to pick one of them, but I found that there was no grip at all for these things, so the first time a model slipped out of my hand. Goldberg described this shape as a hostile shape, and he believes that if his own database can handle these shapes, robotic agility can be better than human hands. For this reason, the Dex-Net database specifically developed an algorithm that attempts to fetch 1000 times in 1000 different ways for each virtual object in the database. Three months later, I visited the laboratory again, where I met Goldberg's proud Jeff Mahler, who now runs the database and has completed the connection of the industrial robot YuMi to the database. “Industrial robots are good at doing repetitive work, but in the face of changing circumstances, robots need to constantly adapt to the new environment they feel. This is a huge challenge,†said Goldberg. With Alexa, Mahler lets the robot put those weird 3D print models in boxes. After the robot touched the item that gave me the horse, it slipped. However, mistakes can also generate new experiences. If you can get hundreds of tests together, you can find tips to grab this item. A robot learns and all the connected robots learn. Amazon also has its own set of robots. In 2015, the e-commerce giant launched the Amazon Robot Contest. The winning robots may enter the shipping center in the future and completely replace human workers. In 2016, the winner of the competition was Delft University from the Netherlands. Their robots removed 12 shipments from large bales and placed them in separate boxes. The robot uses suction cups when picking up smooth-surfaced goods, while others use machine claws. Although the whole process is accurate, the speed is too slow. Berkeley University has not yet participated in the Amazon competition, but this year they will participate in the historic Home Economics robot contest. During the game, the robot has to complete the task of vacuuming, sending meals or cleaning the room. However, there are more restrictions on participating robots. Each team can only use Toyota's human-support robots or Soft Silver's cute pepper Pepper. So what is the attractiveness of home-grown robots? "If you spend $2,000 to make the house tidy and clean, I will not hesitate to buy one." Goldberg predicted. Such robots can not only deal with bear children who throw things, but also help the handicapped or the elderly to do housework. In the future, they may also take on the task of going shopping. Lei Fengwang noticed that a company called Seven Dreamers has already made a laundroid robot Laundroid. After many years of research and development, this product will be officially launched in March this year, but it is very slow, and it will only do a stack of clothes. . Cornell's light pipe gives robots a clever hand. Looking back at the 1973 original Western World movie, the robot installed only slightly deformed hands. However, even after decades, the current robots have made little progress. The YuMi robot developed by Goldberg's own laboratory has two rigid fingers that can be opened and closed like a white shark. If they can use full hands, grabbing items will certainly be much easier. However, the main task of the current Goldberg team is to improve the existing industrial robots. To create the robot's smart hands in the future, the method used may be completely different. In this regard, the Cornell Organic Robotics Laboratory has a certain lead. Their robots have five fingers like humans. These five fingers are made of silicone instead of rigid metal. “In a nutshell, each finger is like a balloon, and the driving force is compressed air. Its principle is similar to our common paintball gun. The bottom of this artificial finger can hardly move, but after it is filled with air, it The top can be curved inwards to simulate human hand movements.†In theory, you can even successfully shake hands with Cornell's robots, and this technology can be used to create bionic hands in the future. With a soft finger, the process of grabbing an item is much simpler. “Our bionic hand can be deformed according to the shape of the object, so it can grab any object without the aid of an algorithm,†said Shepherd, an expert at Cornell University. Cornell's bionic hand is not unique in the industry. Bebonic and Open Bionics have already completed highly skilled dexterous robots. However, they still require humans to operate. Only the upper part of the handicapped disabled limb can be collected. To the precise electrical signal. In addition, the cost of building these robots is enormous, and most people cannot afford them at all. Next, the goal of Open Bionics is to reduce the price of its own robotic products to around US$3,000. Shepherd is very optimistic about its own silicone bionic hand, and once it is mass-produced, it costs as little as US$50. However, the true killer of Cornell University is still on the sensor. They are very accurate and the volume production is quite simple.
Cornell University implanted three polyurethane tubes in each finger of the bionic hand. Researchers called it a light pipe. They can work like fiber optic cables. LEDs and other photodetectors are installed at each end of each light guide. The light passing through the light guide gradually becomes darker as the finger bends. Then, by integrating the data read by the broadcast detector, the robot can obtain the position of each finger and the degree of contact with the object. As the external pressure can be perceived, the fingers can still feel painful in the future. The long road is long. If you want to have the same agility as a human hand, there are many mechanical and computational problems that robots need to overcome. “The human hands are very complicated. They are the evolutionary products of cell-level precision. Our current work is only a low level of imitation. In the future, the number of sensors on the bionic hands of Cornell University may increase from 3 to 100, but in order to achieve We can use thousands of sensors for the density of nerves that rivals ours,†Shepherd said. It should be noted that simply adding sensors does not solve the agility problem of bionic hands perfectly. "We are advancing slowly in the processing of sensor data," Goldberg said. "So, there is not enough light to have a sensor, and the algorithm that goes with it has to keep up. And to tell the truth, we are far from the ideals." For housewives who are troubled by housework, this is bad news. However, it has saved many jobs that people rely on to eat.
It is true that machines have been in our lives for centuries, but the work they can do is rather limited. "They are all placed in fixed locations and do all kinds of mechanical tasks repeatedly," Goldberg said. However, once out of the factory, the machine becomes helpless in unstructured environments such as chaotic rooms and busy warehouses. Grab the trick “Taking an object sounds simple and humans can do it easily without even thinking, but for robots, this action is subtle and elusive,†says Goldberg. But if you carefully consider this process, it really depends on a very complex network in our brains. Take the mug, for example, the human brain will automatically calculate how to hold the cup the most stable, and it will be clearly defined where each finger should be placed. Through human evolution, the brain has its own highly customized processing conventions. "Although I had never seen the pen on the table, I knew I could easily pick it up," Goldberg said. “In the process of taking the pen, the brain regained a similar experience and passed it back to both hands.†Now, Goldberg is learning this trick with his students. To this end, they have specially built a network database called Dexterity Network (Agile Network), where about 10,000 3D virtual items are stored. The scale of database storage of virtual items may gradually expand to a million. When I visited Goldberg's laboratory in September last year, he put a lot of grotesque 3D printing models in front of me. Goldberg asked me to try to pick one of them, but I found that there was no grip at all for these things, so the first time a model slipped out of my hand. Goldberg described this shape as a hostile shape, and he believes that if his own database can handle these shapes, robotic agility can be better than human hands. For this reason, the Dex-Net database specifically developed an algorithm that attempts to fetch 1000 times in 1000 different ways for each virtual object in the database. Three months later, I visited the laboratory again, where I met Goldberg's proud Jeff Mahler, who now runs the database and has completed the connection of the industrial robot YuMi to the database. “Industrial robots are good at doing repetitive work, but in the face of changing circumstances, robots need to constantly adapt to the new environment they feel. This is a huge challenge,†said Goldberg. With Alexa, Mahler lets the robot put those weird 3D print models in boxes. After the robot touched the item that gave me the horse, it slipped. However, mistakes can also generate new experiences. If you can get hundreds of tests together, you can find tips to grab this item. A robot learns and all the connected robots learn. Amazon also has its own set of robots. In 2015, the e-commerce giant launched the Amazon Robot Contest. The winning robots may enter the shipping center in the future and completely replace human workers. In 2016, the winner of the competition was Delft University from the Netherlands. Their robots removed 12 shipments from large bales and placed them in separate boxes. The robot uses suction cups when picking up smooth-surfaced goods, while others use machine claws. Although the whole process is accurate, the speed is too slow. Berkeley University has not yet participated in the Amazon competition, but this year they will participate in the historic Home Economics robot contest. During the game, the robot has to complete the task of vacuuming, sending meals or cleaning the room. However, there are more restrictions on participating robots. Each team can only use Toyota's human-support robots or Soft Silver's cute pepper Pepper. So what is the attractiveness of home-grown robots? "If you spend $2,000 to make the house tidy and clean, I will not hesitate to buy one." Goldberg predicted. Such robots can not only deal with bear children who throw things, but also help the handicapped or the elderly to do housework. In the future, they may also take on the task of going shopping. Lei Fengwang noticed that a company called Seven Dreamers has already made a laundroid robot Laundroid. After many years of research and development, this product will be officially launched in March this year, but it is very slow, and it will only do a stack of clothes. .
Cornell's light pipe gives robots a clever hand. Looking back at the 1973 original Western World movie, the robot installed only slightly deformed hands. However, even after decades, the current robots have made little progress. The YuMi robot developed by Goldberg's own laboratory has two rigid fingers that can be opened and closed like a white shark. If they can use full hands, grabbing items will certainly be much easier. However, the main task of the current Goldberg team is to improve the existing industrial robots. To create the robot's smart hands in the future, the method used may be completely different. In this regard, the Cornell Organic Robotics Laboratory has a certain lead. Their robots have five fingers like humans. These five fingers are made of silicone instead of rigid metal. “In a nutshell, each finger is like a balloon, and the driving force is compressed air. Its principle is similar to our common paintball gun. The bottom of this artificial finger can hardly move, but after it is filled with air, it The top can be curved inwards to simulate human hand movements.†In theory, you can even successfully shake hands with Cornell's robots, and this technology can be used to create bionic hands in the future. With a soft finger, the process of grabbing an item is much simpler. “Our bionic hand can be deformed according to the shape of the object, so it can grab any object without the aid of an algorithm,†said Shepherd, an expert at Cornell University. Cornell's bionic hand is not unique in the industry. Bebonic and Open Bionics have already completed highly skilled dexterous robots. However, they still require humans to operate. Only the upper part of the handicapped disabled limb can be collected. To the precise electrical signal. In addition, the cost of building these robots is enormous, and most people cannot afford them at all. Next, the goal of Open Bionics is to reduce the price of its own robotic products to around US$3,000. Shepherd is very optimistic about its own silicone bionic hand, and once it is mass-produced, it costs as little as US$50. However, the true killer of Cornell University is still on the sensor. They are very accurate and the volume production is quite simple. Cornell University implanted three polyurethane tubes in each finger of the bionic hand. Researchers called it a light pipe. They can work like fiber optic cables. LEDs and other photodetectors are installed at each end of each light guide. The light passing through the light guide gradually becomes darker as the finger bends. Then, by integrating the data read by the broadcast detector, the robot can obtain the position of each finger and the degree of contact with the object. As the external pressure can be perceived, the fingers can still feel painful in the future. The long road is long. If you want to have the same agility as a human hand, there are many mechanical and computational problems that robots need to overcome. “The human hands are very complicated. They are the evolutionary products of cell-level precision. Our current work is only a low level of imitation. In the future, the number of sensors on the bionic hands of Cornell University may increase from 3 to 100, but in order to achieve We can use thousands of sensors for the density of nerves that rivals ours,†Shepherd said. It should be noted that simply adding sensors does not solve the agility problem of bionic hands perfectly. "We are advancing slowly in the processing of sensor data," Goldberg said. "So, there is not enough light to have a sensor, and the algorithm that goes with it has to keep up. And to tell the truth, we are far from the ideals." For housewives who are troubled by housework, this is bad news. However, it has saved many jobs that people rely on to eat.
Laptop Computer,Core I5 Laptop,Metal Slim Laptop,Fhd Laptop
C&Q Technology (Guangzhou) Co.,Ltd. , https://www.gzcqteq.com