Ackerman, Evan. “Dextrous Robotics Wants To Move Boxes With Chopsticks.” IEEE Spectrum: Technology, Engineering, and Science News, IEEE, 13 July 2021, 15:44 GMT, spectrum.ieee.org/automaton/robotics/industrial-robots/dexterous-robotics-develops-chopstick-manipulation-for-boxes.
Article title: Dextrous Robotics Wants To Move Boxes With Chopsticks
This approach could be more versatile than grippers or suction
Screenshot: Dextrous Robotics
My summary of the article:
Dextrous Robotics, a startup led by Evan Drumwright – a former professor at GWU – and his ex-student, has put forth a new approach to box handling robots. Dextrous Robotics identifies that conventional box-handling robots (or any object-manipulating robots) utilize suction systems, and that although suction systems are indeed advantageous in many situations, it lacks versatility (or adaptability to changing environments and circumstances). For example, a robot with a suction system finds trouble manipulating small objects, especially if that object is smaller than the suction itself. Given that the size of suction cups is fixed at a relative small size in order to be able to lift heavy objects, this is a serious problem when it comes to versatility. Moreover, a suction system is very sensitive to external environmental variables such as temperature, moisture, and smoothness of the interacting surface (small particles like dirt may affect the performance of suction cups).
Dextrous Robotics comes up with a very creative solution in order to resolve such problems associated with the conventional suction robot; a chopstick robot. The startup company believes that the greatest advantage of a chopstick robot over anything else is, as opposed to a suction robot, its high versatility. Similarly to real chopsticks, a chopstick robot can adapt to and function in a lot of circumstances, rather there be variables in the environment or in the object itself that the robot is trying to lift.
The company realizes that a sophisticated software is just as necessary in developing a successful product as is a great hardware. In line with this realization, the company has developed a simulation that they could use to test the software of the robot without actually implementing the software in the real world with the real robot. This way, the company could greatly shorten the time it takes for the chopstick robot (both hardware and software) to be developed, since the simulation allows for the software to be developed, updated, and checked just as a real world implementation would do. And even though the company realizes that their simulation is not perfect (they believe the simulation agrees with reality about 95% of the time), the fact that the simulation is 95% accurate gives them great confidence that it can indeed be used to check the software of the robot.
My response to the article:
Although I have seen many iterations of industrial robots for many different uses, one that uses chopsticks sounds unprecedented to me. Being Korean and having used a pair of chopsticks for all my life, I am very confident that chopsticks are indeed an exceptionally versatile tool. There is practically nothing a chopstick cannot lift up, since it applies force in two directions and an expert chopstick user can control the strength of both forces very effectively and individually. Given such specifications, I believe there is no reason for a chopstick robot (of course, with a good software that enables the robot to adapt in different circumstances) to be ineffective in the workplace.
Personally, I enjoy making simulations a lot. These are some simulations I made myself (Link1, Link2). And I believe this article has a great point about the importance of simulations in many areas (including product design, as used in the production of Dextrous Robotics' chopstick robot). Simulations offer an environment for people to test a product or an idea in the virtual world before they do so in the real world, and I believe the greatest advantage gained from this is the saving of time and money. Simply put, doing something in the real world as opposed to doing so in the virtual world is much more time consuming and expensive, considering the equipments that have to be bought and the setup that needs to be done – let alone paying the people who conduct such experimental works. Reading this article has further solidified and extended my interest in making simulations, since I believe that one day the skills I have built in this area will pay off.
コメント