Critical systems are those in which a system failure can cause irreparable economic or human damage. In this type of system, the correct behaviour of the system must be guaranteed before it is launched. In order to carry out this process, the system must be accurately characterised, and a thorough study of its behaviour in the worst case scenario must be carried out. System characterisation processes describing the system components, their interactions and internal behaviour are of vital importance, as well as the analysis and verification systems that allow to check the interactions between the components to ensure that they will meet the system requirements in the worst case scenario.

Critical systems are part of larger complex systems in many application domains, such as industrial control, automotive systems, aerospace systems, etc.

Cyber-physical systems are the evolution of embedded systems towards an architecture with greater connectivity. In industrial and transport environments, for example, such systems tend to have a large number of critical components, given their strong interaction with the physical world. However, this greater connectivity, which gives them the flexibility that makes these technologies truly disruptive, is what adds the most complexity to their development. The system components must not only be correct from a functional and security point of view, but also from the point of view of their temporal behaviour and data integrity. Although some of the time constraints were already present in the old embedded systems, the presence of non-critical or highly variable computational components on the same platform adds a level of uncertainty that must be properly addressed.

These systems are the building blocks of promising initiatives in the field of industry, such as Industry 4.0, the Internet of Industrial Things (IIoT), Digital Twins (DigitalTwins), Industrial Fog Computing architectures, etc. All of them are strongly related to each other to the point of being, in many occasions, concepts that cannot be approached in an isolated way. In all of them, cyber-physical systems are the basis on which all these initiatives and architectures are built.

Cyber-physical systems are those requiring both a strong interaction with the real world and the use of general purpose data networks in order to carry out their mission. A series of time restrictions are usually imposed when carrying out this interaction in such systems. This is due mainly to their interaction with the environment. Such systems are generally called real-time systems, and the time at which system responses occur is as important as the responses themselves. Real-time systems require guarantees prior to their implementation, which ensure that the system will behave correctly when deployed. These guarantees, often imposed by the certification processes, are based on a modelling of the system components and the interaction between them, an exhaustive characterisation of the behaviour of each of these components, and the analysis of the behaviour of the entire system in the worst possible scenarios or conditions. All this guarantees that, once the system is in operation, it will behave within the established parameters.

The components and/or functionalities of the system are usually represented by a set of recurrent tasks, and the results of the analysis are those that assure that these tasks are going to have a correct temporary behaviour. However, the use of general purpose data networks and associated components introduce a certain level of uncertainty, both in the network load and in the less critical components needed to handle it. The presence of multiple levels of criticality on the same execution platform introduces greater complexity in both the analysis and implementation of such systems.

In addition, real time systems in the real world usually show changes in functionality, depending on the phase of their mission—e.g. the phases of a flight: taxi mode, take-off, climb, cruise mode, etc.—or the presence of failures in the system that can temporarily or permanently degrade its behaviour, be it external failures or unexpected behaviour of some of its components. Such systems are called multi-mode systems, since they can behave differently depending on the way they are. Since the functionality of a system is represented by a specific set of recurring tasks, a mode change involves the execution of a different set of tasks. In such systems, system modelling and analysis must consider the correct behaviour of the system, both in the stationary states and in the transitions between the different operating modes. This implies, as mentioned previously, the activation of certain tasks in the new mode and the deactivation of some tasks in the previous mode, which should not be executed in the new mode.

The integration of intelligent components that give cyber-physical systems a certain autonomy and adaptability to change, as well as the uncertainties that this adds into the predictability of the system’s response in its interaction with the physical world, is one of the greatest challenges that arise regarding cyber-physical systems. Maintaining precise and reliable behaviour while remaining flexible is one of the goals that these building blocks must address in the coming years, as they are integrated in the cognitive processing mechanisms that will be added to the new industry. In the same way, the development of modelling, analysis and simulation environments and tools that facilitate their fast and reliable integration into the different application domains is of the utmost importance.


Get in touch with us through the form for companies and we will guide you to incorporate these technologies into your project through the partners specialized in your activity.

Selecciona idioma: