Using artificial intelligence to achieve greater effectiveness and efficiency in car handling | Tests at one of the world's largest car ports | Project partners BIBA – Bremer Institut für Produktion und Logistik at the University of Bremen, BLG LOGISTICS and 28Apps Software

Bremen, Bremerhaven, Berlin. Approximately 70,000 car parking spaces, around 1,000 car ships per year, plus thousands of car trains and trucks: with 1.7 million cars handled annually (2022), the BLG AutoTerminal Bremerhaven is one of the largest in the world. This is precisely where the tests in the "Isabella" and "Isabella 2.0" research projects took place. They dealt with the process planning and control of automobile handling in ports with the help of mathematical optimisation, simulation and artificial intelligence (AI). “Isabella 2.0" has now also been successfully completed. The partners involved were BLG LOGISTICS, 28Apps Software (Bremen) and BIBA – Bremen Institute for Production and Logistics.

The range of services at the BLG terminal includes almost all vehicle logistics services, from storage and technical preparation of the cars to their handling. The planning and control of the highly complex and dynamic processes of car handling in inland and seaports can fundamentally be made more effective and efficient with the "Isabella" system, according to the results of the research. The system offers options for increasing the number of transshipments, reducing ship lay times and making better use of space - and thus securing competitiveness and jobs.

Support from the terminal

According to project partner BLG, the solutions developed are promising for the future and their application in practice is currently being discussed. An important factor had also been that the terminal staff had already been involved in the project during the prototype tests of the system. Experiences of the operational users on site were incorporated into the R&D work, and occupational psychologists were also involved.

Rapid adaptation to current conditions

The intelligent planning and control system focuses on the movements of cars in sea and inland ports and integrates the time-critical processes of the external transport carriers train, truck and ship with their loading and unloading. With the help of mobile data acquisition and real-time status messages, the control algorithm enables individual assignments of driving orders in real time, route optimisation for the driving personnel shuttle buses and thus rapid adaptation to current conditions.

An interactive, digital interface visualises the terminal area three-dimensionally: a multi-touch table shows the image of the virtual model of the terminal, its digital twin. This supports planning and control and leads to better results. Different levels of detail and all relevant planning information such as terminal occupancy can be displayed on the table. The Isabella system can evaluate different planning scenarios based on simulation and display the results on the table.

Mobile apps for easy use

The assignment of orders is digitised, the assignment of orders for vehicle movements on the terminal is done depending on the location of the vehicles as well as the individual drivers. A control algorithm was developed for this and tested within a simulation environment. In the real system, communication between the control system and the staff at the car terminal takes place via mobile apps. To determine the exact locations of the drivers and vehicles, research was conducted into suitable positioning systems. In addition, reliable data reception is needed everywhere. To achieve this, the project partners used the new 5G mobile radio standard in trains and trucks, for example, as well as a local communication network for data receppage 2 of 2 tion in ships. Here, a prototype mesh WLAN was evaluated. It has good response times, transmission speeds and ranges throughout the hull of the ship.

What was not possible before

For the logistical performance of the system, sensitivity analysis methods were used to determine the ideal ratio of car drivers on the terminal to shuttles. An optimisation algorithm was developed for the allocation of driving jobs to drivers and shuttles, which can also map the operational uncertainties of a terminal. A classic optimisation algorithm usually requires a lot of computing time because many options have to be run through. A neural network leads to results much faster. Therefore, in a second step, a new AI approach was used to create a more computationally efficient representation of the optimisation system. For the first time, optimisation problems including constraints can be mapped. This allows the fastest possible adjustments to current situations.

Virtual reality supports staff

The training environment and applications developed in the project use virtual reality (VR), for example via VR glasses. Among other things, new driving personnel can be trained on various scenarios of their future tasks, including virtual journeys through the terminal. Other training units address common sources of errors and answer questions such as where to store vehicle keys in the many different car models on the terminal.

Practice-oriented interaction of research, development and application

In the two projects, BIBA conducted research on the planning and control of an automobile terminal under operational conditions and, in particular, developed a powerful optimisation algorithm for the control of vehicle movements. As an application partner, BLG LOGISTICS contributed its practical knowledge along with data and provided the test environment for testing with the Bremerhaven, Hamburg and Kelheim car terminals. After the end of the project and in further cooperation, developments from the Isabella projects could now be used here in productive operation. Technology and development partner 28Apps-Software developed the software solutions for the virtual training environment, the multi-touch table and the technical implementation of the optimisation algorithm to implement the concept.

Key data on the "Isabella" and "Isabella 2.0" projects

The research and development work, which lasted a total of 6 years, comprised two projects: "Isabella" ("Automobile logistics in sea and inland ports: Interactive and simulation-based operational planning, dynamic and context-based control of device and load movements") and "Isabella 2.0" ("Automobile logistics in sea and inland ports: Integrated and user-oriented control of device and load movements through artificial intelligence and a virtual training application”). The two Isabella projects had a total budget of 7.3 million euros and were funded by the Federal Ministry of Digital Affairs and Transport (BMDV, previously BMVI) in “Innovative Hafentechnologien” (IHATEC) programme a total of 5.1 million euros, as well as being accompanied by the project management organisation TÜV Rheinland.

 https://isabella2.de
 https://projekt-isabella.de
 www.youtube.com/
 www.biba.uni-bremen.de
 www.blg-logistics.com
 www.28apps.de

Prof. Dr.-Ing. Michael Freitag (Leiter BIBA), Telefon: +49 421 218-50 002, E-Mail: fre@biba.uni-bremen.de

Marit Hoff-Hoffmeyer-Zlotnik, M. Sc. (BIBA), Telefon: +49 421 218-50 094, E-Mail: hhz@biba.uni-bremen.de

Britt van Delden (Presse BLG), Telefon: +49 421 398-34 75, E-Mail: Britt.van-Delden@blg.de

Emrah Gencer (28Apps Software), Telefon: +49 421 301 59 900, E-Mail: e.gencer@28apps.de

• 

  AutoTerminal Bremerhaven1 Foto BLG-HaukeDressler.jpg

  AutoTerminal Bremerhaven1 Foto BLG-HaukeDressler.jpgCars that have been unloaded or are to be loaded are mainly transported by rail. The AutoTerminal Bremerhaven has 16 railtracks and head ramps. Image: BLG/Jan Meier
• 

  Isabella Abscannen1 Foto BLG-JanMeier.jpg

  Isabella Abscannen1 Foto BLG-JanMeier.jpgCommunication between the control system and the staff at the car terminal takes place via mobile apps. A special location system was developed in the Isabella project to determine the vehicle loca-tions. Image: BLG/Jan Meier
• 

  Isabella 199 Foto BLG.jpg

  Isabella 199 Foto BLG.jpgMichael Görges, Isabella project manager of BLG LOGISTICS, demonstrates the multitouch table in test use at the BLG AutoTerminal. Different simulation-based planning scenarios can also be pre-sented via this medium. Image: BLG/Sönke Denis

You can download the following images and others in print quality summarized as a ZIP file here: F&E-Projekte zur intelligenten Planung und Steuerung.zip