KAnIS: cooperative autonomous intralogistics systems

-The KoopLogistik subwork package is dedicated to the central question of what new possibilities are opened up for the optimization of intralogistics processes through the targeted networking and cooperation of individual fleet members. Research is being conducted into new strategies for order planning and processing, predictive maintenance, and energy and battery management for industrial trucks. The common feature of these strategies is the targeted evaluation of a data set that is as extensive as possible, provided cooperatively and fused in the cloud. Artificial intelligence methods, the latest network theory approaches (e.g., from graph theory), and discrete-event simulation methods are used here.

Important objectives of the optimization of intralogistic processes are:

• the increase of the internal flow of goods and materials as well as the reduction of goods and material throughput times
• the recording of the maintenance and wear condition of the individual fleet members
• Based on this, the realization of intelligent order planning and predictive maintenance of the FFZ members (predictive maintenance) to prevent unnecessary downtime
• Reducing energy costs by integrating innovative energy management systems into autonomous and learning logistics systems.
• Maximizing the uptime of fleet members by implementing an intelligent battery/charging management system and an innovative inductive battery charging process for electrically powered FFZs.

In this subwork package, a novel cooperative approach to automating in-plant industrial truck fleets is being researched. In this approach, the individual industrial trucks (FFZs for short) make the data from their environmental sensors, their position information, and other important vehicle data available to a central unit, the cloud, via a communication platform (see the sketch below). Conversely, to supplement the data from the on-board sensors, each member of the fleet is also provided with important sensor information and vehicle data from the other FFZs, as well as the information from other sensors stationarily mounted on the premises, via radio transmission from the central unit. By means of a suitable sensor data fusion, all vehicles, regardless of their individual sensor equipment, have access to almost the same information about the vehicle environment (cooperative perception) and their own position (cooperative localization).

Another focus of the KoopAutonomie research project is the development of methods for automatic lane guidance of FFZ. Innovative control algorithms are being researched which, on the basis of the data provided by cooperative perception and localization, enable targeted control of steering, drive and braking and thus ensure a highly accurate target path sequence (see the following sketch) as well as adherence to the specified target speed course.

Special attention is paid here to the strongly varying steering behavior of the individual fleet members caused by the different FFZ variants and variable vehicle parameters (e.g. variable vehicle mass and center of gravity position due to load absorption). It is imperative that the latter be taken into account in controller design in order to ensure stable control loop behavior and thus safe vehicle guidance. The aim of the research is to develop a control concept that can be used universally for all FFZ variants and that adapts itself independently to the changing FFZ variants and vehicle parameters by using state-of-the-art machine learning methods.

Prof. Dr.-Ing. Bochtler, Measurement and Circuit Technology
Prof. Dr.-Ing. Doll, Cooperative Automated Traffic Systems
Prof. Dr. Eley, Logistics and Optimization
Prof. Dr.-Ing. Krini, Signal and Sensor Processing
Prof.-Dr. Möckel, Mathematics, Machine Learning, Predictive Maintenance
Prof.Dr.-Ing. Mußenbrock, Energy Management and Efficiency
Prof. Dr. Stark, Mathematics, Signal Processing, Project Coordination
Prof. Dr.-Ing. Teigelkötter, Electric Drive Technology, Battery Management
Prof. Dr.-Ing. Thielemann, Microsystems Technology and Sensors, Biological Neural Networks
Prof. Dr.-Ing. Zindler, Automation and Control Technology, Vehicle Control