Implementation and demonstration of concurrent task execution for OPC UA skills

OPC UA is a widespread standard for data exchange and semantic modelling of data and interfaces across different industry sectors. In previous work, a generic skill concept based on OPC UA has been developed that can be used to access hardware and software functionality via a standardized interface and connect different modules to a higher-level process control system, facilitating increased flexibility and reduced commissioning time for production systems. In this approach, each skill can be executed once at a time, which is suitable for most applications. However, in some cases, the order and resource allocation for the task execution is mandated from the overall process control to an external resource, such as a feet controller. In this case, the process control is required to add its tasks to a task queue which is then handled by the external resource, while still keeping track of its state of execution.
In this work, we present an addition to the current skill implementation that enables concurrent task scheduling and execution, which is required for tracking tasks that need to be delegated to other task controllers or that can be executed simultaneously. We demonstrate our approach by providing an interface to the Open Robotics Middleware Framework (Open-RMF), which controls a feet of autonomous, mobile robots. Tasks, such as transport requests, are generated by the high-level process controller and then submitted to the task queue of Open-RMF via our standardized OPC UA interface. This allows for a consistent and continuous integration of process control with OPC UA for a wide spectrum of use cases.