- Amazon robots managed 224 units / hour, compared to humans at 243 units / hour
- Robots are more consistent, but humans can be faster in some cases
- Additional research is necessary to prevent damage
After a series of robotics announcements in recent years, including its latest Vulcan robot which has a sense of touch, Amazon tests new generation models called “ Stow ” and “ Pick ”, and their names say everything – they are designed to help in the storage and selection processes in the realization centers.
However, the company has recognized that current robots technology is not in place to replace human workers despite promising results.
According to his performance tests, humans were on average 243 units per hour compared to robots, at 224 units per hour. The difference is small, but Amazon also highlighted certain shades.
Amazon’s robots are good, but not human good
Where the robots stand out, it is consistent. “It has also been found that humans had a greater variation in rupture rates: people can quickly store many small items effectively, but are slower with large objects, crouching for lower bins or when using a step ladder to reach the upper bins,” said Amazon.
The company has also observed how humans can do several tasks at the same time with two hands, such as removing a storage tank from one hand and storing one element with the other – something that the current generation of robots does not do.
However, the gap ends, Amazon’s robots defeating success greater than 85% at a rate similar to humans on more than half a million tests.
That said, there are challenges in the implementation of robotics through Amazon warehouses, including the damage caused by the abolition of products, the insertion of products in storage bins and the crushing of light boxes.
As such, although complete human replacement is not yet possible, hybrid systems are very effective, with more work necessary to improve the handling and reduction of damage caused by robots.
“More in -depth research on predicting the interaction of complex elements may be necessary for the scale of more general manipulation,” added Amazon researchers.
