Collaborative project: Dynamic behavior analysis of driverless transport vehicles in predefined road networks based on log and layout data (SwarmLogiX)
Duration: 01.07.2025 to 30.06.2027
Funding reference: 100772464
In the "SwarmLogiX" research project, together with our project partner FlowLogiX GmbH, we are pursuing the goal of fundamentally improving the planning, simulation and optimization of complex production and intralogistics systems in the semiconductor industry.
Modern production environments, known as fabs, are highly dynamic overall systems in which production and transportation are closely interlinked. While subsystems such as production planning or manufacturing execution systems (MES) have already been considered in detail, there has been no consistent simulation of the overall system to date. In particular, the behavior of transport systems (OHT) is insufficiently mapped in existing models due to inadequate control logic. As a result, key optimization potential remains untapped, for example in terms of resource utilization, throughput times or adherence to delivery dates.
This is precisely where SwarmLogix comes in: The project is developing a data-driven, learning approach for automated rule recognition, modeling and simulation of complex transport systems in the context of the overall factory.
The project pursues three core objectives:
- The automated analysis of real movement and transport data in order to identify behavioral patterns of transport systems and derive resilient decision rules.
- The creation of a hybrid simulation approach that combines rule-based methods with machine learning processes in order to realistically depict the behavior of the overall systems.
- The continuous improvement of the models through iterative feedback, so that the system behavior can dynamically adapt to new conditions and data situations.
A key result of SwarmLogiX is a new type of simulation and analysis platform that considers production and transport systems in a common model space for the first time. This makes it possible to visualize complex interactions in the overall system and make well-founded statements about optimization potential.