Collaborative work between workers and drones in a factory environment

Humans and robots have long interacted and collaborated in factories. So what happens when these robots fly?

Collaborative robotics has transformed the modern economy, enabling rapid transportation and increasing productivity in various manufacturing industries. But what happens when robots working alongside humans are replaced by drones whizzing by?

In the near future, drones will be of great use in many environments. They will be able to move shipments and products from one end of a logistics warehouse to the other and have general material handling functions. Space that is currently useless in factories and warehouses may become available flight paths. Drones are fast and can participate in existing production activities without costly modifications.

Inevitably, they also carry a certain amount of danger. When a drone hits a worker, it can easily kill someone. Likewise, accidents such as dropped packages caused by drone malfunctions can have fatal consequences. Effectively identifying and avoiding obstacles in time, especially those who move quickly and irregularly, will be the main challenge for drones in warehouse applications.

This is where motion capture technology comes in. With the help of this technology, manufacturers can visualize and capture reference data, which is crucial for guiding human activities and drone tracking.

As a company specializing in motion capture systems, Vicon recently established a research team in cooperation with the Technical University of Dortmund. The German-based university is working to change the way drones, robots and people operate safely in warehouse environments.

The research project is based at the Innovation Lab of the Technical University of Munich, an interdisciplinary research center equipped with a Vicon motion capture system. About 40 Vicon motion capture cameras are installed in this system, and the captured data can be analyzed with corresponding software.

In one brilliant demonstration, a man in a motion-capture suit walks through a swarm of 12 autonomous drones. It's a horrifying sight, but the drone's ability to avoid the testers with unerring precision demonstrates the drone's ability to react in seconds, thanks in large part to the data captured by the cameras. With a motion capture speed of up to 300 frames per second, the camera can simultaneously process large amounts of data and respond quickly with high precision.

This is a demo on how to make motion capture a redundant system for the drone's on-board sensors. The information captured by the cameras is used to safely deploy drones or robotic reference systems in warehouse environments and to check that the sensors on the drones are working properly.

On a research level, motion capture and onboard machine vision allow developers to give machine learning programs the ability to predict erratic human movements.

The lab also employs machine learning techniques that allow drones to recognize movements such as carrying boxes and delivering goods from one place to another. This enables the development of safety protocols and gives them flexibility so that future drones will not require fixed infrastructure to navigate.

"Our work has profound implications and benefits for the manufacturing and logistics industries," said Christopher Reining, a research associate. "We plan to organize this data to ensure that other researchers can access and use it."

All of this is aimed at enabling drones to work alongside humans in dense warehouse environments. If you think this seems unlikely, remember: Many in the robotics industry believed that robots would never be able to work alongside humans outside of cages. But now this situation is very common in industries such as logistics and light manufacturing.

Author: Greg Nichols

This article is from a translation. If you want to reprint, please obtain authorization from this site first.