LPDL: A customised language to document load profiles

Abstract - Digital twins not only reflect the current state, but also anticipate future developments, simulate scenarios and offer valuable insights for the optimisation of real development processes. In order to design these complex digital models efficiently and precisely, however, we need powerful tools that fulfil the special requirements of these virtual worlds. This is precisely where the ‘Load Profile Description Language’, or LPDL for short, developed in TwinSpace comes in.

Challenges of software development in the automotive industry

Software development in the automotive industry frequently faces challenges, as it often occurs independently of the subsequent integration into the overall system. This results in runtime tests only being carried out at the end of the development process. If one discovers only then that the developed software does not meet the requirements, this can lead to considerable costs. More powerful processors are often used to avoid cost-intensive redevelopment, but this has a negative impact on costs, energy consumption, and therefore CO2 emissions.

The role of the LPDL in solving this problem

The LPDL is a textual description language for mapping real load behaviour per component. As part of TwinSpace, it serves as an interface for the various tools developed in the project. By extracting load profiles from C-code, the XML-based language can map non-functional properties of software, such as memory consumption or processor load. The load profiles can either be taken from existing legacy projects or provided based on the suppliers’ current requirements – without disclosing confidential information. In addition, the LPDL offers the option of gradually refining the load profile by iteratively integrating empirical values from the development process, thus successively providing a finer and more precise profile. This process takes place in parallel to the V-model. The LPDL maps the essential control flow as well as the type and number of instructions without modelling specific calculations or algorithms. Overall, the LPDL contains all of the important information for further processing and serves as input for the twin generator, for example.

Perspectives for software development

In TwinSpace, the LPDL is initially being tested in the automotive sector, although the underlying language is agnostic to the actual application and can be employed across industries. This opens up a wide range of potential applications in different sectors in the long term. The development of LPDL to optimise the resource requirements of embedded systems is not only a technical challenge but also offers the opportunity to fundamentally revolutionise the way we work with virtual models and simulations. In an increasingly digitalised world, this language could offer a decisive advantage in making complex systems more efficient, flexible and intuitive. The road ahead is ambitious, but the potential benefits are enormous and could hold the key to the next big innovation in digital twins. The selection of toolchains and hardware based on a digital twin can be done in the early stages of development to match the actual load of a system, avoiding oversized hardware and saving not only power during operation but also valuable and rare resources.