Towards a general framework for design synthesis and development of mechatronics and cyber-physical systems

The fourth industrial revolution, which fosters technological advancement, has given rise to a new product development trend based on the design synthesis philosophy, known as Mechatronic System Design. The synthesis of these increasingly powerful and complex products is a difficult proposition as it involves integrating a large number of elements from many engineering disciplines, with strongly coupled physical energy fields, having little or no cross-functional understanding of emergent system behaviour. Cyber-Physical Systems (CPS) are a new breed of complex, hybrid mechatronics systems; coined in 2006 by the US National Science Foundation, CPS is characterized by concurrent, dynamic and real-time interactions among heterogeneous subsystems from cyber and physical domains. CPS and mechatronic systems are best designed as a whole as opposed to combining individual subsystems sequentially as design failures are often caused by interactions between the subsystems. A general purpose framework to help with the integration of subsystems, architectural exploration management and design automation is unavailable and must be developed. A popular software engineering approach called Model-Driven Development (MDD), which focuses on automatic software production and reuse allows designers to automate and integrate new functionalities using a model-centric approach, thus simplifying and formalizing design tasks. However, it has three major drawbacks: Inability of current frameworks to support complete design automation causes implementation gaps; current frameworks support signal based discrete-time models, while mechatronic models are more accurately represented as hybrid and continuous-time models; and existing frameworks cannot accommodate new, evolving methodologies like VDI2206. This work aims to extend the promise of MDD approach to automate the design of Mechatronics and CPS, through the development of a general framework for design synthesis. In the first part of this work, a framework for rapid application design and mechatronic system development has been developed using software engineering and IT approaches. Using services as functional building blocks, complex systems were built as loosely coupled components and services; this approach helps software architects evolve flexible design solutions, system designers improve modularity and system integrators understand and workaround limitations of concurrency and composition. While the service architecture appeared practical and viable, physical and hardware especially the real-time communication and concurrency issues remained. To overcome this, a formal Model of Computation and coupling between systems at energy levels was required. The second part focuses on the development of an alternative framework that combines cyber and physical models using an energy-based approach; no metamodeling frameworks employing energetic approaches are currently available for design synthesis. A unified modeling framework has been developed using Bond Graphs (BG-UMF); it is vendor neutral and involves all design phases, from concept-to-code and implementation. It allows simulation of the system at all stages, allowing early detection of problems. A methodology for design synthesis of mechatronics systems has also been developed, bridging the concept-implementation gap. The framework is also modular, allows design reuse and enables automatic model transformation and code generation, taking designers a step closer to complete design automation of mechatronic and CPS.