A large portion of the technical systems surrounding us implement software executed by a “computer.” The failure of these systems can have disastrous consequences on financial, environmental, or human levels. Therefore, their development is subject to strict constraints aimed at lowering the probability of such events to a level deemed acceptable from both economic and social perspectives.

The “Critical Embedded Systems” Center of Competence (CSEC) at IRT Saint Exupéry’s mission is to participate in defining these constraints and providing the necessary tools to meet them.

IRT Saint Exupéry invites you to explore a series of articles dedicated to this highly strategic research topic. In this first article, we offer an overview of this research area. We will highlight some of the significant results achieved by IRT Saint Exupéry over the past ten years in this field.

Embedded systems and critical real-time embedded Systems: what are we talking about?

Real-Time Systems

Did you know? An embedded system is a hardware and software combination designed to control a larger system, such as a car or an aircraft. It typically interacts with its environment and must meet the time constraints imposed by that environment. It is referred to as a “real-time” system because it must react based on “real” time, meaning “physical” time.

Time can be measured in various ways. For example, by the ticks of a clock or by other physical phenomena, such as the rotation of an automobile’s engine shaft to a specific position. Regardless of the chosen measurement, a real-time system must provide a service before a precise deadline, or even at a given deadline. Any failure to meet this constraint significantly reduces the value of the service and is considered a failure.

Critical Real-Time Systems

Did you know? A system is generally classified as critical when its failures can have catastrophic consequences on its environment, whether technical, financial, ecological, or human.

Many embedded systems are critical. These include flight control systems or autopilot systems in aircraft, or speed regulation systems or airbag control systems in automobiles. In these examples, system failures can lead to serious consequences, affecting both people (passengers, people on the ground, or near the vehicle) and the environment.

Hide this software that I can’t see…

Real-time embedded systems automatically process information from physical measurements (such as speed, temperature, pressure, etc.) or from devices that allow interaction with a human operator (for example, the position of the steering wheel). These are computer systems, where the role of the software is paramount. In this context, the complexity of the software naturally increases with the complexity of the functions performed by these systems.

To assume that the methods and tools used to develop these software are fully satisfactory, mature, and perfectly mastered is a mistake. The history of computing is far from over. Software production remains, even today, an intellectually complex and, above all, critical task, as a software failure can lead to a functional breakdown. Moreover, the activities of design, implementation, and verification of software represent an increasingly significant portion of the cost of system development. This raises questions about our ability to master “software cathedrals” that are becoming taller and more fragile. Thus, research activities in the field of software are essential, as evidenced by the considerable number of research teams working on all aspects of software development.