On intelligent automation systems

Sammanfattning: Developing automation systems that are capable of handling dynamic and unpredictable situations is a challenging task, as it requires adapting to a changing environment and managing potentially unforeseen action outcomes. In contrast to traditional automation, where control code is explicitly pro- grammed, a model-based approach might be a more appropriate solution for automating such systems. Such an approach allows for integrating planning algorithms, which can enable the generation of control sequences that consider the system’s state. This capability is essential in enabling human-robot col- laboration and handling error recovery and restart. We refer to such a model- based and goal-oriented approach to automation as Intelligent Automation Systems (IAS). To bridge the gap between research and practical utilization, this thesis aims to facilitate the development of IAS by investigating methods for their preparation, control, and testing. A framework for preparation and virtual commissioning of IAS is presented, which compiles the necessary methods into a high-level structure, aiming to streamline the IAS development process. As part of the preparation process, an effort to explain the unsolvability of some planning problems by localiz- ing potential faults in behavior models is presented. Furthermore, this thesis investigates planning and SAT solving methods aimed at improving the effi- ciency of planning, thereby enhancing the responsiveness and adaptability of IAS. A planning and execution framework for IAS is presented, with a focus on handling dynamic and unpredictable systems. Finally, an iterative method for the verification of IAS is presented, where methods such as supervisory con- trol theory, model checking, unit and integration testing, and property-based testing play key roles in ensuring the correct behavior of IAS. Connected to verification, a criterion for assessing the test coverage of IAS is presented. This research contributes to the field of intelligent automation by providing solutions for the development, control, and verification of systems designed for complex and unpredictable environments, aiming to bridge the gap between theory and practice.