From production to warehouse to shipping – wherever goods are manufactured, sorted, or packaged, picking is also involved. In this process, robots often take individual goods out of crates and put them into new groups. In the FLAIROP project, Festo is conducting research with partners from Germany and Canada to increase picking robots’ intelligence using distributed AI methods. They are exploring how to leverage training data from multiple stations, production facilities, or companies without having to disclose sensitive company data.
“We are exploring how to use the most multifaceted training data possible from multiple sites to develop more robust and efficient solutions using artificial intelligence algorithms than would be possible with data from just one robot,” explained Jonathan Auberle of the Institute of Materials Handling and Logistics (IFL) at the Karlsruhe Institute of Technology (KIT).
As part of this, autonomous robots process items by gripping and moving them at multiple picking stations. The robots are trained with totally different items at the various stations. Ultimately, the goal is for them to also be able to grip items from other stations that they have not encountered before. “By using distributed learning, also known as federated learning, we are able to strike the right balance between having a wide range of data available and keeping data secure in the industrial environment,” said Auberle.
To date, federated learning has primarily been used for image analysis in the medical sector, where protecting patient data is obviously of paramount importance. This is why an exchange of training data, such as images or gripping points, does not take place when training the artificial neural network. Instead, only pieces of stored knowledge – the local weights of the neural network that show how strongly one neuron is connected to another – are transferred to a central server. The weights from all of the stations are collected there and optimized with the help of various criteria. The improved version is then fed back to the local stations and the process repeats itself.
The aim is to develop new, better-performing algorithms for the robust use of artificial intelligence for industry and Logistics 4.0 while complying with applicable data privacy guidelines.
“In the FLAIROP research project, we are developing new ways for robots to learn from each other without sharing sensitive data and company secrets. This offers two major benefits: we protect our customers’ data and we become faster because the robots can perform many tasks more quickly. For example, the collaborative robots can help production workers complete repetitive, difficult, and tedious tasks,” explained Jan Seyler, Head of Advanced Development, Analytics, and Control at Festo.
“DarwinAI is delighted to make our Explainable (XAI) platform available to the FLAIROP project and pleased to work with such esteemed Canadian and German academic organizations and our industry partner, Festo. We hope that our XAI technology will enable high-quality human-in-the-loop processes for this exciting project, which represents an important facet of our offering alongside our novel approach to federated learning. Having our roots in academic research, we are enthusiastic about this collaboration and the industrial benefits of our new approach for a range of manufacturing customers,” commented Sheldon Fernandez, CEO of DarwinAI.
“The University of Waterloo is thrilled to be working with the Karlsruhe Institute of Technology and a global industrial automation leader like Festo to bring the next generation of trustworthy artificial intelligence to manufacturing,” said Dr. Alexander Wong, Co-Director of the Vision and Image Processing Research Group, University of Waterloo, and Chief Scientist at DarwinAI.
“By leveraging DarwinAI’s Explainable AI (XAI) and federated learning, we can create AI solutions that help factory workers perform their daily production tasks, thereby increasing efficiency, productivity, and safety.”