Release date: 2016-05-31
At present, robots still can't do something, one of them is that they can't feel pain. It is for this reason that we can put robots in some dangerous environments, or let them do things that humans are not willing to do or that threaten human life. However, two German researchers from the University of Hanover in Germany believe that in some cases, feeling pain and responding to pain may be a good ability for the robot.
The researchers developed a "artificial robotic nervous system" that teaches robots to feel pain, and in order to avoid damage to their motors, gears or electronic components, these robots need to respond quickly to pain. Last week, they presented the project at the IEEE International Conference on Robotics and Automation (IRA) in Stockholm, Sweden.
Why is it a good idea to let the robot feel pain? In the case of researcher Johannes Kuehn, the answer to this question is actually the same as the answer to "why let humans feel pain is a good idea." “Pain is a system that protects us,†he said. “When we avoid the source of pain, we can help us not be hurt.†Those who can't feel the pain are often more likely to get hurt because their body is affected. The damage can't make an instinctive response in time.
Kuehn is working with Professor Sami Haddadin to develop the project. Professor Sami Haddadin is one of the most famous physics and human interaction and security experts in the world. Kuehn said that by protecting robots from damage, their systems can also protect humans. We know that the number of robots is growing. In many cases, they work close to human workers. Some undetectable damage to robotic equipment can lead to production accidents and pose a threat to the safety of human workers. Kuehn and Haddadin concluded that if our human biological mechanisms can feel and respond to pain so effectively, why not design a bionic robot controller that mimics the biological mechanisms of pain? This controller responds to pain reflexively and then protects the robot in potentially dangerous interactions.
In fact, the idea of ​​this robotic reflection controller is not a new concept. About five years ago, Torsten Kroeger and his colleague La Sapienza at Stanford University and the University of Rome developed a similar controller to help the robot arm avoid collisions with humans.
Watch this video, which shows that the controller allows the robot to avoid collisions with humans, but these devices were originally designed to prevent the robot from hitting humans, but did not consider the safety of the robot itself (except for some attached products). In fact, designing an artificial intelligence robotic nervous system that can sense pain and respond to pain is a completely different matter. It needs to be based on how humans respond to painful touches, as Kuehn and Haddadin wrote in their ICRA papers:
Robots need to be able to detect and distinguish between unpredictable physical states and disturbances. For those situations that may cause potential harm to themselves, the robot needs to be rated and take appropriate countermeasures, ie reflections. In order to achieve this goal, the human prototype has become our inspiration, which means that the human reflection of pain will be used to design the robot pain sensory model and response control. Based on an understanding of human pain research, we focus on "normalizing" the pain perception of a robot as an explanation for touch.
This video shows how a controller prototype runs on Kuka's arm. This controller is equipped with a “bio-tactile fingertip sensor†that senses pressure and temperature. The author seems to be able to feel the sound of a robot that makes "hey, it hurts!", and as the pain becomes more and more intense, its voice will become higher and higher every time:
This robotic haptic system uses a "neural robot tissue model inspired by the human skin structure" that determines how much pain the robot should feel when subjected to a certain magnitude of force. Just like human neurons, when the force exceeds a certain threshold, the model will repeatedly transmit the peak of pain information. At this time, the pain controller will respond according to the pain information classification. The pain classification includes mild, moderate, and severe. pain.
At the [mild] pain level, this type of contact may cause the robot to be injured, or to prevent the pain from occurring when the robot is performing a task. The robot "feels" uncomfortable and should retract smoothly until the end of the contact. A strong collision is a [moderate] pain level. The robot "feels" moderate pain and should quickly retract the operation away from the source of the pain until the end of the contact. After that, the robot can choose to return and terminate the operation. [Severe] The pain level contains all the contact that may cause damage to the robot. At this time, the robot may need “help†to avoid pain. In order to prevent this damage from causing more damage, the robot should switch to gravity compensation with additional damping to dissipate the pain and improve the safety of the robot and the environment in which the robot is used through its strict passive behavior.
In terms of biological perception and control, Kuehn and Haddadin's paper first explored a pain-based robotic reflection controller. Although this situation sounds like a sense of disaster, the technology benefits most robots, not only Improve robot autonomy and ensure the safety of robots and human workers when working together. Of course, ensuring that robots don't harm humans is a top priority, but if we want robots to appear safely around us, we must first make sure that the robots are protected from harm.
Source: AI Technology Review
Fruit juice drinks is a beverage made of fruit juice products obtained by physical methods such as pressing, centrifugation and extraction. Fruit juice retains most of the nutrients in the fruit, such as vitamins, minerals, sugar and pectin in dietary fiber. Often drinking fruit juice can help digestion, moisten the intestines and supplement the lack of nutrients in the diet.
There are two kinds of fruit juice: pure fruit juice and compound fruit juice. As far as pure fruit juice is concerned, it can be divided into two types: one is to directly use the original juice, and the other is to dehydrate the freshly squeezed fruit juice first, and then add water to reduce it into concentrated fruit juice. Most pure fruit juices are concentrated. AGOLYN offers various flavors of honey tea, such as honey Citron tea etc.
Juice Drinks,Honey Tea,Honey Citron Tea,Goji Berry Juice
Xi'an Gawen Biotechnology Co., Ltd , https://www.agolyn-bio.com