The increasing complexity of logistics, which among other things consists of changing customer requirements, a high variety of variants and the demand for fast reaction and delivery times, requires highly flexible and automated logistics and production. Due to the large amount of digitally available information and approaches from the field of artificial intelligence, automated transport systems can be designed more intelligently with the help of novel models for environmental perception, thus reducing the costs of intralogistic material flow. Due to growing cost constraints, manufacturing companies are increasingly demanding the automation of non-value-adding activities. These processes are becoming more and more complex and demand automation, which is particularly evident in changing boundary conditions, complicated chaining of events and increased security requirements. An example of this is the internal material flow, which is to be realized in the future more and more by autonomous transport systems. Especially promising are developments in the field of artificial intelligence and robotics, which should enable the change from automated to autonomous transport systems.
The aim of the dissertation project is the development of a semantic environment model for autonomous transport systems using artificial intelligence and machine processing of knowledge.
A transport system in intralogistics can access both its internal sensors (e.g. depth camera or laser scanner) and the external data already available in the logistics environment (e.g. ERP system or control system). In order to be able to react optimally to changes and events in the environment, the environment must be perceived and understood. The first step is to obtain information from the sensor data using special semantics. The resulting information can then be linked to knowledge. The entirety of the knowledge about the environment then makes it possible to derive decisions automatically. The result of this work enables an intelligent system behaviour of transport systems in intralogistics through the development of a generic model for the acquisition and processing of data from the digital and real environment for intelligent, intralogistic transport systems. Furthermore, the system enables the communication of structured data from the environment of the transport system back into the cloud. In addition to material flow, fleets of transport systems can also perform other functions such as automated inventory or process monitoring.
The research project can be divided into the following work packages. First, the state of the art of research and technology is elaborated and analysed by means of literature and market research. This is followed by the requirements analysis for the target system: both the logistics-specific requirements are discussed through expert discussions and analysis of the transport processes, and technological requirements are set for the model. This information is then compiled and evaluated with the help of a "Requirement Engineering" process in the form of a requirement specification and existing defects are revised again. With the help of the present requirement specification, the development of the generic semantic environment model then begins. The V-model is used for this purpose: First, the development of the concept of the basic architecture begins. Subsequently, the data and information sources in internal logistics are analyzed in detail. The main question to be clarified here is which internal and external data and information is generally available to the transport system in logistics. Based on this, appropriate modules will be created which enable the extraction of information from chaotic data and thus a perception of the environment. The available information is then combined temporally and across several sensors. The information is finally structured by means of methods of the mechanical processing of knowledge before conclusions can be drawn about it. Finally, the demonstratory implementation of the developed system and the evaluation of real logistics processes at the BMW Group take place.