Dextrous robots might be with us very soon
Robots and computers are ridiculous, as a matter of fact! Very ridiculous. Any change in the working environment outside the pre-programmed process can eventually shut down the entire production line. Here, improving and enlightening robot skills offers an incredible opportunity for a higher degree of flexible automation.
In robotics, the word “dexterity” is often used. Dexterity refers explicitly to the skill of performing tasks by hand. It is used by robot makers who describe their robots as dexterous. Robotics researchers use it to explain the dexterity of their development. The robot has the potential to dive into existing workflows and quickly deliver significant value. But not in all areas of the industry. However, one of the areas in which some robotics companies have recently entered is the handling of boxes. Specifically, it unloads the container box from behind the truck much faster than humans. For robots, it is an excellent activity that demonstrates its strengths. You can work in a semi-structured environment where speed, energy, and accuracy are highly valued, and it’s not a job that humans are particularly interested in or trying to do. According to
Statista: In 2019, approximately 75,000 logistics robots were sold worldwide to meet the increasing efficiency requirements of global trade relations. The value will continue to increase and should exceed 250,000 by 2023. Dextrous Robotics, a startup led by Evan Drumwright based in Memphis, Tennessee, is one of the latest approaches to accomplishing such tasks. Initially, Evan Drumright was a professor at George Washington University. After that, he spent several years at Toyota Research Institute. In 2019, Drumwright co-founded Dextrous Robotics with his ex-student Sam Zapolsky. They have developed a strategy to operate the box without suction or grippers. Instead, use a pair of movable arms best described as chopsticks that hold the crate in place.
The idea of using a movable arm like chopsticks is excellent! However, the reason for using chopsticks is that you can grab any item, no matter how big or large it is. This is useful in tight spaces with significant shape, size, and weight differences. They are also suitable for operation and can push and place objects accurately. Initially, Dextrous focused on trailer unloading work. The additional maneuverability means that more complex maneuvering tasks can be considered in the future, such as loading onto trailers. This happens as often as unloading, but it isn’t very easy for automated robots.
Every coin has two sides! The technique of skillfully using the chopsticks arm has both advantages and disadvantages. The biggest drawback of the chopsticks technique is that it is much more challenging to use such an approach.
In the meantime, using suction systems in handling the crate is very consistent, which Dextrous has observed so far in other companies in this field. Drumwright explains that due to limited trial and error, the suction system works quickly and reliably when there is a flat, impermeable surface (like most boxes). However, the suction system is very sticky and has limited handling. This means that each task can be difficult and time-consuming to perform correctly. Other suction issues include the sensitivity to temperature and humidity, the tendency to pick up as much dirt as possible, and the fact that the suction device needs to be designed based on the most prominent and heaviest items expected to be treated in the future, included. The final problem with the suction system is that if you want to work with small objects, you have far more options than you would like.
Manipulating small objects in the suction system is a challenge! However, that does not mean that the intake system is always inferior. Drumwright agrees that an extraction system is viable for specific professional jobs. At the same time, if the chopsticks are operated well, they will be more versatile.
The secret of a dexterous chopstick robot is to focus on stimulation. Dextrous finds it difficult to crack the hardware. They want to create something that works only once, rather than repeatedly. Getting it right the first time is not practical, but the closer you get, the more you can make suggestions in advance.
Dextrous can predict how your system will perform even on the most challenging tasks to simulate, such as contact and stiction. This is possible because very loyal hardware modeling, real-time simulators, and software can convert directly between sim and real. Ultimately, Dextrous’s goal is to create a system that can complete complex tasks faster than other methods.
In the future, we may program the robot to symbolize seeing and grabbing something, but existing research is trying to teach the robot how to do this through trial and error.
SUGANDHA GUPTA is the writer of this article. Views expressed and information provided belong solely to the author.