Boston Dynamics Turns Humanoid Robot Atlas Into Perfect Warehouse Clerk

About a year ago, Boston Dynamics introduced the all-electric Atlas robot, which is faster, stronger, and more agile than its hydraulic predecessor. After a year of experimentation and training, the company found one of the most suitable jobs for a humanoid robot – work as a storekeeper on an assembly line. The robot fits perfectly into the Hyundai car assembly plant team and is ready to go to work for other companies.

Image Source: Boston Dynamics

Sorting parts for feeding to the assembly line is a routine, monotonous, but important and quite complex task. Parts must be sorted into cells in the right order depending on the assembly of specific models of cars on certain days of the week. In addition, different models can be assembled on one line, and it is important to select components for each of them. At the same time, there are a lot of components, which requires a good memory and the ability to quickly find the necessary parts in the warehouse.

For a robot, memorizing information and orientation in space are not a problem, but it is difficult for the machine to adapt to a constantly changing environment. Components arrive at the warehouse in deep containers, and the robot cannot always look into all corners and examine the parts from different angles. In order to get parts from the warehouse cells and place them in a container for feeding to the conveyor, the machine must be able to squat, bend, stretch and manipulate objects of different weights and dimensions.

It is easier for a person to cope with all these operations, but when the work is reduced to endless repetition of the same type of actions, it becomes a serious burden on the psyche. The Boston Dynamics team trained the electric Atlas to work as a storekeeper for about a year and achieved certain success, as can be judged by the presented video.

It should be noted that the task of teaching the robot the skills of a storekeeper had to be solved with relatively simple programming tools. Using powerful AI programs and neural networks in this case would have been an excessive solution. In order to give the robot logistics skills, the company developed relatively simple algorithms, breaking down complex tasks into a chain of simpler actions. As a result, the robot showed itself well in servicing Hyundai assembly lines, and Boston Dynamics admits that it has already received requests for similar “iron workers” from other companies.