So far, small haulage companies and transport service providers in Germany have been little digitized. The planning and organization of journeys are usually done with pen and paper, communication mainly by phone or fax. »Up to now, this has also worked smoothly for us. But the demands are increasing: We must meet tighter and tighter schedules and become more and more flexible,« says Jens Markowski, Managing Director of ecs-Logistics in Kreuztal in the Siegerland region. With around a dozen employees, the freight forwarder is one of the many companies in the industry for which digitalization has not seemed absolutely necessary until now. However, says Jens Markowski: »Even in view of the increasing orders, we are now exceeding the limit up to which manual scheduling still works.« In addition, the communication effort between the dispatcher in the company, the customers and the drivers makes it difficult to concentrate on the actual planning tasks. Within this context, the forwarding and logistics company has now made itself available as a pilot partner for the new »AI-based ETA« solution.
Company benefits from neutral development
The arrival time prediction software was developed as part of the Silicon Economy project at the Fraunhofer Institute for Material Flow and Logistics IML. The mission of the Silicon Economy is to develop a digital infrastructure for the platform economy of the future. All software and hardware is developed open source, i.e. it is freely available – a basic principle of the Silicon Economy. This opens up the opportunity, especially for less digitized and small and medium-sized enterprises, to push ahead their digitization, because they can simply place the software on their IT infrastructure – usually with the support of an IT service provider – and integrate it into the system environment. The »KI-ETA« software was initially developed without major industry participation, although in exchange with potential users, in order to develop a product that is as neutral as possible for the market.
With the implementation project of the Mittelstand-Digital Zentrum Ruhr-OWL, the software developers can now enter the proof-of-concept. »We are pleased that we can now test the solution in a company and, above all, develop it further,« says Kai Hannemann, research assistant at Fraunhofer IML and member of the »KI-ETA« project team. »We expect to be in a position to enable our project partner to digitally plan its transports more efficiently and sustainably within a short period of time.«
Tracking solution is integrated
The researchers are setting up a dashboard for the Kreuztal-based freight forwarder that can be accessed via a classic website on the Internet. This is hosted by Fraunhofer IML, i.e. the company does not need any infrastructure of its own. At the same time, the transport service provider receives an encapsulated system environment to which only the company has access. All data circulate only within the company and remain there. Via the dashboard, the ecs-Logistics dispatchers can plan their tours, create new transports and calculate routes and driving times with the help of AI.
For the first time, the »Silicon Economy« experts will also connect the software with a tracking solution. A corresponding app will be installed on the drivers’ mobile phones. The dispatchers can then track where which truck is at any time on the computer. The DGSVO-compliant solution only runs when the trucks are moving at a certain speed. The drivers also have the option to switch off their mobile phones completely at any time. »At the beginning, we will certainly do both and continue to plan the transports manually as well. However, as soon as the solution is stable, the goal is to switch over completely,« says ecs-Logistics Managing Director Jens Markowski.
System gets smarter through real data
Through the piloting, the Silicon Economy experts at Fraunhofer IML are gaining real data for the first time and can thus improve the self-learning forecasting system of their solution. The routing is based on road data from the independent Mobility Data Marketplace (MDM) for North Rhine-Westphalia. This data is gathered, for example, via measuring posts on motorways. Information about traffic regulations, closures or even about restrictions on overtaking for trucks is passed on every minute and fed into the »KI-ETA« system. Up to now, however, the truck guideline speed of 80 kilometers per hour has been used to calculate arrival times. However, experience has shown that this speed is not always reached everywhere – for example in rush hour traffic or on steep inclines. »As part of the implementation project, we now receive real data directly from the truck which teaches the artificial intelligence in our system,« Kai Hannemann is pleased to say. »In principle, this works like the voice-controlled digital assistants at home: every input makes the software smarter – or in our case: every real piece of information makes our forecasts even more accurate.«
The implementation project started back in February and will run for a total of six months.
- Added value at a glance
- Display of all active transports as an overview in a table
- Tracking of the current position of individual transports
- Determination of the arrival time forecast of a transport on the basis of truck speeds
- Solution based exclusively on open data sources
- AI improvement possible through specially recorded tracking data
- Data misuse excluded through »encapsulation« of the system