DARKO

Safe, functional and efficient robots for logistics in Factory 4.0

In order to meet the ever-changing conditions and demands of the market, the manufacturing industry worldwide is moving towards more agile, efficient production. The agility of production depends crucially on the effectiveness of intralogistics - simply put, the movement of parts and objects around the factory. Logistics robots have the potential to change the system, but only if they have the necessary capabilities. They need to be highly flexible, cost and energy efficient and able to work safely and smoothly with humans in shared environments.

The DARKO project therefore aims to develop a new generation of agile production robots to close the "skills gap" where today's robots fail. We are developing energy-efficient elastic arms that can pick up and throw objects in a human-like way. Importantly, we are also exploring how robots can operate safely in unknown, changing environments, in part by sensing humans and their intentions in order to interact with them smoothly and intuitively. Our robots also need to be easily deployable in new factories to reduce the cost and effort currently required to map a new site, establish routes and traffic rules, etc. Finally, the DARKO robots will have predictive planning capabilities to take the most efficient actions while limiting risks.

Objectives of DARKO

Our overarching goal is research and innovation for efficient and safe intralogistics robots in agile production. Specifically, we want to realize the fundamental technologies that enable flexible, robust and easy-to-use intralogistics in agile production. Our work focuses on the following five topics:

• Handling efficiency: Today's industrial robots can grasp objects, but mostly in a static and stationary environment. DARKO will enable more advanced handling capabilities through novel hardware that is safe and energy efficient, better computer vision for recognizing difficult objects and agile control and perception for throwing. 

• Efficiency in human-robot co-production: For intralogistics robots to work efficiently, they must coexist with human personnel. We want to increase efficiency and safety by learning the robots' action patterns so that they can adapt to them, and by predicting human intentions and finding new ways to communicate the robots' intentions. 

• Efficient deployment: One of the main obstacles to the market launch of robots today is the deployment effort. We will facilitate robot deployment from the perspective of fail-safe mapping and localization as well as the "self-completion" of robot maps with existing maps of different types.
• Risk-aware operation for safety and efficiency: We incorporate risk assessment as a driving principle for the decisions made by the robots, including the overall orchestrated behavior of robots, humans and machines.
• Demonstration of feasibility in an integrated system: Finally, a central topic is the practical demonstration in an integrated system of how the above-mentioned scientific objectives can be implemented. ARENA2036 is the perfect environment for this, in which we will simulate a use case based on real requirements, which will serve to demonstrate and validate our scientific results. 


This is how the DARKO project is structured

The DARKO consortium brings together a unique combination of partners with complementary, world-leading expertise. In this way, the requirements for robots that work together with humans in intralogistics applications of agile production can be addressed. A number of key technologies are being researched and further developed in seven scientific work packages as part of the project:

• Novel elastic manipulators and flexible end effectors: The project partner Technical University of Munich (TUM), which builds on globally recognized expertise in the field of safe manipulation, is leading the work on the development and control of a novel and efficient elastic manipulator in work package WP1. The project partner, the University of Pisa, has extensive experience in the field of manipulation and in particular the design of end effectors. It will contribute to the development and control of new types of flexible end effectors.
• 3D perception and scene understanding: As part of WP2 (3D Perception and Scene Understanding), partner Bosch will work with Örebro University, EPFL, the University of Pisa and the University of Lincoln to conduct research into more robust, efficient and intelligent AI-based methods. This will enable mobile and industrial robots to gain a better semantic understanding of their static and dynamic environment. To this end, multimodal data from lidar and vision-based sensors will be combined with supervised, weakly supervised and self-supervised deep learning techniques.
• Mapping and localization: The coordinating partner, Örebro University, will lead research in mapping and localization technology to further improve the robustness of existing SLAM systems and enrich them with higher level information.
• Safe dynamic manipulation: The University of Pisa is leading work package 4, which is concerned with researching new methods for safe and dynamic manipulation. The project partner École Polytechnique Fédérale de Lausanne (EPFL) will contribute to the manipulation and throwing tasks in WP4 with its extensive experience in the field of robot throwing. The Technical University of Munich will focus on safety-related aspects, while the University of Örebro will work on the combination of vision and grasping.
• Human-robot interaction on a spatial level: The University of Lincoln with its Lincoln Centre for Autonomous Systems (L-CAS), together with the University of Örebro and Bosch, will research more efficient approaches for social navigation in the presence of humans as part of the WP5 work package.
• Motion planning: Bosch is leading the work package WP6 (Predictive Motion Planning). The aim is to research methods for planning and controlling mobile robots on wheels in dynamic environments. Predictable human conditions and possible (safety) risks will be taken into account. The University of Pisa and the Technical University of Munich will make an important contribution to this task with their existing background knowledge of safe and risk-aware planning.
• Risk management: The project partner ACT Operations Research IT is leading the technical work package 7 (risk and planning), in which a framework for risk assessment and prediction is being developed together with the University of Pisa to ensure both efficient and safe operation in joint human-robot environments.

DARKO receives funding from the European Union's Horizon 2020 research and innovation program under grant agreement no. 101017274.