Continuous delivery of innovative solutions at low risk
Dr. Martin Benedikt
Group Leader Co-Simulation & Software
The automotive industry is facing a radical change. Exploding system complexity is motivated by alternative (electrified) powertrains, automated driving or the introduction of innovative and highly-connected technologies in general. Handling the number of product variants and facets naturally leads to an immense effort in system design, development and validation. In addition, there is a significant pressure to deliver robust and innovative technical solutions on short term. With established development methods, this problem is difficult to counteract.
Computer-aided approaches like CAE or numerical simulation enable a system analysis prior to the development of expensive prototypes. Combining the geometric, functional and dynamical product features leads to a digital prototype for frontloading purposes. This Digital Twin can be used to exploit the systems at hand to reduce expensive redesign cycles in later development stages. In connection with parameters and measurement data approaches like predictive engineering and predictive maintenance ensures further potential for reducing costs and effort, also after development during operation.
Otherwise, safety-critical and secure systems make it mandatory to provide functionally proved solutions which motivates for Model-based Systems Engineering (MBSE) – an approach where collaboratively created domain models are exploited as the primary means of information exchange between engineers, enabling consistent system designs.
Indicated Trend and Solution
Model-based Systems Engineering covers the full system development cycle from design to validation, had its beginning 30 years ago and struggled with argumentation on the return-of-invest (ROI), i.e. system modelling effort versus its benefit. Nowadays, systems are that complex and are engineered by broadly distributed teams (x-company) where the ROI on model-based Systems Engineering gets obvious – a clear trend indicates the settlement of MBSE as key for efficient system development.
MBSE is understood as a tool supporting the overall system development process. Having the MBSE principle and the related toolchain in place it’s a matter of steering and orchestrating the system development from a more abstract point of view via the processes on top. For ensuring high-quality development processes efforts are taken in the automotive industry (Automotive SPICE, ISO 15504) to improve and to assess software development processes.
Especially within initial or (radically) new system design phase’s classical system development processes, such as the V- or the waterfall-model, could be enriched by agile development cultures, which has been proven for decades in software development. Trends indicate the adoption of the agile Continuous Integration & Continuous Delivery (CI/CD) paradigm for development of mechatronic and cyber-physical system, where software, simulation models and first hardware components mounted on testbeds are continuously and automatically integrated, tested and approved for deployment. This approach ensures a truly consistent development status besides allowing a flexible adaptation of requirements, reducing development risks and subsequently leading to transparent, secure and accelerated system development.
The combined and integrated approach of MBSE and the agile manifesto would unleash synergy effects and significant, still hidden potentials for taking bold steps towards efficient development with the ultimate goal of having: continuous delivery of innovative solutions for mechatronic and cyber-physical systems at low risk and effort.
The establishment of a full-blown CI/CD tool-chain requests related processes, methods and tools to be available and in place. Without neglecting the experiences done in agile software development, some fundamental-research and application related, crucial topics has to be addressed first:
- missing link between system-level modelling and system science
- purpose-driven system modelling and design
- CI/CD tool-chain for mechatronic systems
- agile enrichment of classical development processes
- processes, method and tools for truly consistent system development approaches
Application-oriented actions clearly support the transition as well as the transfer of knowledge from research to pre-development stages within companies. With its long-term experience in modular and cross-domain system development the VIRTUAL VEHICLE Research Center works together with German OEMs and Suppliers on the realization of future MBSE-based CI/CD approaches. Impressive proof-of-concept demonstrators for Virtualization as well as CI/CD Tool-Chains are already established within the K2DM (K2 – Digital Mobility) Living-Innovation-Lab, fostering the transition from research to application.