The use of cyber-physical systems has increased dramatically in recent years. These systems consist of embedded computation components that interact with the physical world, and are found in, for example, cars and other vehicles, manufacturing equipment and consumer products. They have applications in the automotive, aerospace, telecommunication, robotics, and healthcare industries.
Many of these systems, such as automotive electronics and medical equipment, are safety-critical, with stringent requirements for reliability and speed. Safety-critical functions in vehicles include, for example, braking, active safety, and autonomous driving features. Also other functions in vehicles, such as image and speech recognition, intelligent navigation, and sensor fusion, require very high computation capacity and guaranteed quality. The increased complexity of all these functions places high demands on the processors and system architecture. We are facing the challenge of how to design complex heterogenous architectures that satisfy the requirements placed on such a system, both functional and non-functional. These requirements concern the hardware, architecture, software and control. Advanced optimisation is needed here to satisfy all of the design constraints.
LiU is home to more than 25 years’ experience of design at various system levels, and in the development of optimisation methods for various types of cyber-physical systems. The techniques that have been developed can also manage both transient and permanent errors, intrusion and other threats, process variations, and problems that arise during large variations in temperature. The technology can cope with not only dynamic changes in workload, but also major disturbances such as hardware failure and malicious attacks.
Methods and tool are also developed for system specification, formal verification, testing and design for testability, and combined design of hardware and software. Research is also carried out within systematic design and design automation techniques for industrial applications.
Zebo Peng, Professor Software and Systems. Photo credit Anna Nilsen