In the future, there will be robots that can lead teams in factories. Traditional industrial robots are efficient but rigid, only follow procedures and cannot cooperate with humans. Now, a team of human and robot partners, and the robot is responsible for the command of the team in the PK victory over the human expert group. In the future, this model is expected to greatly improve production efficiency
robot + human vs human team
if an engineer forms a team with a robot and challenges a team of experts with more than 10 years of welding experience, who can weld the metal frame of the body of a Hummer (an off-road vehicle) faster? Which side do you think will win
the engineer who teamed up with the robot is Michael Dawson Haggerty. After receiving his master's degree, he entered the Robotics Institute of Carnegie Mellon University as an engineer
his competition partner is a robot. Different from the common robot with no matter what kind of plastic packaging, this 4-meter-high welding robot with a mechanical arm called spitting fire has been improved, equipped with a laser that can act as an eye, and can work side by side with humans
in addition, the work is not ordered by Dawson Haggerty, and the cooperation of this two person team is more like another way: first, according to who can complete a task more effectively, Dawson Haggerty and Spitfire will allocate those difficult positioning and welding work, and the robot will command the process. Among them, robots, rather than humans, often make decisions
that is to say, in this competition, Dawson Haggerty is led by robots
after such division of labor and cooperation, Dawson Haggerty and his robot partner spent 10 hours and $1150 (including raw materials and labor costs) to weld the frame; As a control group, welding experts spent 89 hours to complete the same work, and charged $7075
defects of traditional industrial robots
in the past half century, manufacturing enterprises have been using industrial robots to achieve automation to improve production efficiency, and they work very well. Although robots have always been used to perform specific repetitive operations (such as welding a specific set of interfaces for each car on the production line), humans have carefully arranged and set up the production line to fill FM - the total average friction force in the experiment uses the powerful power and high accuracy of the robot man
this mode is very effective in the production of mass-produced products such as automobiles. However, with the increasing demand for personalized or customized production, manufacturing enterprises are increasingly manufacturing small batch products according to user needs, and the time required to set manufacturing processes (such as welding or processing) has become a major bottleneck. Preparing robots for specific jobs takes too long, sometimes months. In the meantime, people must arrange the welding sequence, reassemble components, write programs for robots, prepare inventory materials, and optimize welding parameters
traditional industrial robots only obey the orders of programs, but cannot cooperate with humans. Traditional industrial robots do not give up until they reach their goals according to their nature, no matter whether human beings are in front of them or not. If an obstacle blocks its movement, the industrial robot will enter the wrong state and automatically shut down. Although this situation is better than crushing someone, it greatly reduces the production efficiency. The EU evaluates the safety rate of the general plastic recycling process in food contact materials. If humans and these colleagues just pause when they are too close to each other, how much less work can they delay
robots lead humans
the next generation of industrial robots will be fundamentally changed and can work safely around humans. If a robot accidentally hits a person, its impact force should not be fatal, or it is not enough to bring danger at all. Robots will understand the location of human beings in the plant, and they should be able to use voice, posture, facial expressions, words, graphics and other means to communicate with their human colleagues
taking the combination of Spitfire and Dawson Haggerty as an example, Spitfire can not only learn from human beings, but also be smart enough to guide human work. Spitting fire will divide a large project into small steps, and objectively allocate these tasks according to who (robot or human) can complete them faster, without subjectively giving priority to one party
the frame welding work of Dawson Haggerty and Spitfire starts from extracting the bill of materials from the CAD description file of the car frame. Based on this list, the computer in the robot automatically arranges which parts to order from the supplier and how to cut the standard size steel pipe to the exact length. Then, the computer will plan the best welding process, and if the damage is serious, the spring should be replaced to give the best way to fix the parts, so that they can remain stable during the welding process
Spitfire is also equipped with a small projector, which allows it to project images and words directly onto the frame. These images enhance the representation of entities, and robots can use it to guide Dawson Haggerty step by step to set up complex manufacturing processes. Dawson Haggerty put all the materials and parts in place to do these jobs, which are usually regarded as coolies. Human beings are actually a better choice. Because these relatively light parts have variable shapes, human hands are more suitable to hold than robot arms
flamethrower can also use laser displacement sensors to accurately perceive the three-dimensional workspace, and check to confirm that all components are correctly placed. Using the projector and sensor, it can mark the precise position on the car frame with light, and guide human work in the process of welding and assembly
but after the two partners arrange the parts that need to be welded, spit fire and take over the rapid welding work. Spitting fire is not only a fast welder, it takes only 5 seconds to complete a 5cm welding, and its welding technology is also super first-class. Usually, before starting each welding work, professional welders need to adjust about 20 important welding parameters. The fire spray will test the effect of these trial welds and adjust their own settings for improvement. It can be said that this robot has become a professional welder through self-study
but spitting fire is not perfect. In some cases, the robot can't complete some specific solder joints. At this time, it will guide Dawson Haggerty to help with some difficult work
if human beings can cooperate with robots, factories will no longer need to set up those cumbersome assembly lines in a time-consuming and labor-intensive manner, which can save billions of dollars in equipment installation costs. It's not a problem to improve the best-selling products: human and robot partners can design and manufacture any customized products for customers, from electronic equipment to aircraft, without expensive equipment upgrades. This technology will enable enterprises to quickly respond to user needs, and the product update cycle will be shortened from several years to several weeks. Therefore, we need to realize that in the future, robots may be more efficient in leading work teams rather than just engaging in production
LINK
Copyright © 2011 JIN SHI