SOMC: Service-Oriented Modell-based Control – Dynamic Software for Dynamic Systems

为了提高基于模型的控制的适应性和可维护性，本项目追求采用具有高度内在灵活性的面向服务的软件体系结构。该研究项目的目的是对面向服务的、基于模型的控制进行科学调查、概念开发、实施和实验验证。在数字化和设备日益网络化的时代，控制和自动化技术的主要挑战是(i)能够在离线和运行时对控制系统进行调整，（ii）实现灵活的系统集成，以及（iii）保持日益复杂的控制系统的可维护性。这些挑战很难用最先进的软件架构来解决，因为控制软件通常是基于静态集成的框图设计的。然后将框图的各个功能分配给单独的ecu。由于这种固定结构，控制回路的单个组件很难被替换，因为更改单个功能通常需要修改所有其他组件。因此，控制回路的自适应几乎是不可能的，因为它们意味着不同的信息流方向。这个项目提出了一个解决这个问题的新方法。我们建议将控制回路的元素表述为面向服务的软件体系结构的服务。因此，软件系统中的信息流不再与物理系统的因果方向相对应。预期的优点是：(i)服务的灵活互连（通过使用编排）允许在运行时适应控制系统，以及（ii）服务的良好定义的接口增加了可维护性。在这个项目中，实现面向服务的控制循环的必要步骤是将控制循环的元素系统地表述为服务，开发用于动态集成服务的编排机制，以及评估所产生的动态效果。由于面向服务的软件体系结构的服务对其需求和保证使用类型化接口，因此我们将研究基于模型的控制方法，以便将根植于控制理论的保证与服务领域的保证联系起来。特别地，我们将考虑服务保证和用于控制设计的模型之间的联系。该项目将使用两个不同的测试平台验证方法，这将真实地展示分布式控制系统中出现的现象。因此，测试平台将允许批判性地评估我们的新方法的性能，并将其与经典方法进行比较。

This project pursues employing service-oriented software architectures, which possess high inherent flexibility, in order to increase the adaptability and maintainability of model-based control. The aim of the research project is the scientific investigation, conceptual development, implementation and experimental validation of service-oriented, model-based control.In times of digitalization and increasingly networked devices, the main challenges of control and automation technology are (i) to enable adjustments of control systems both offline and at runtime, (ii) to enable flexible system integration, and (iii) to keep the increasingly complex control systems maintainable. These challenges are difficult to address with state-of-the-art software architectures, where control software is classically designed based on block diagrams that are statically integrated. Individual functions of the block diagram are then assigned to separate ECUs. Due to this fixed structure, individual components of the control loop can hardly be replaced, as changes to individual functions often requires modifications in all other components. Therefore, adaptations of the control loop are hardly possible as they would imply different information flow directions.This project proposes a novel approach to solve this problem. We suggest formulating the elements of the control loop as services of a service-oriented software architecture. Therefore, the information flow in the software system does not correspond anymore to the cause-effect directions of the physical system. The expected advantages are (i) the flexible interconnections of the services (by using orchestration) enable adaptations to the control system at runtime and (ii) the well-defined interfaces of the services increase the maintainability. The necessary steps towards a service-oriented control loop in this project are the systematic formulation of the elements of the control loop as services, the development of orchestration mechanisms to dynamically integrate the services and the evaluation of the resulting dynamic effects. Since services of a service-oriented software architecture use typed interfaces for their requirements and guarantees, we will investigate model-based control methods, in order to link guarantees rooted in control theory to guarantees in the service domain. In particular, we will consider the connection between service guarantees and the model used for control design.This project will validate the methods using two different test benches, which will realistically showcase phenomena arising in distributed control systems. Thus, the test benches will allow to critically evaluate the performance of our novel approach and comparing it with classical approaches.