Research Centre in Real-Time and Embedded Computing Systems

CISTER has well-established roots in the real-time and embedded systems (RTES) scientific community. The unit also fosters its activities well aligned with its national and international partners from both industry and academia. As its strategic vision, the unit has been consistently able to identify and contribute to emerging topics in the area, and continues to do so with a strong tradition of developing foundational work in relevant topics.

With their ubiquitous deployment, embedded platforms are becoming increasingly complex as they grow more powerful, owing to their obligation to address ever increasing and demanding requirements. Their complexity and resource-awareness brings further challenges to the development of reliable and efficient systems, such as resource (e.g., CPU, memory, power) management, novel operating systems and virtual machines and timing analyses.
---
CISTER strategy is to tackle research efforts within proper societal and economic contexts. The general expectation is that real-time and embedded systems will enable response to the three wake-up calls that society has had in recent times: pollution peaks, economic crisis, and societal aging. These developments indicate a need for better use of natural, industrial and human resources, respectively. Real-time and embedded systems are expected to enable better use of resources, with reduced waste and pollution, by providing more (and better) information and more sensitive (and finely tuned) monitoring and controlling in all domains.

From an application perspective, these systems occur at the local level with specific improvements in sectors such as homes, offices, vehicles, factories, traffic management and healthcare, to a ‘scaled-up’ global level such as smart-cities, smart-regions and even smart societies. It is thus important to leverage the technological and scientific results of the centre with application-oriented efforts on specific societal concerns. Our research activities address the following application areas:

- isolation and performance issues in emerging platforms for automotive and avionics sectors;

- large scale dense monitoring and control of smart cities;

- sensor systems for healthcare and ambient assisted living;

- real-time system behaviour in factory automation;

- smart ubiquitous embedded platforms.
---
In particular, with a strategic vision for the future, CISTER is working in emerging topics such as:

- next generation of computing systems' programming paradigms;
Next generation of Real-Time Embedded Systems will be based on heterogeneous parallel architectures, and will need to consider multiple dimensions of complexity such as performance, energy efficiency, time-criticality, dependability, and cost-effectiveness. We research new programming paradigms that allow increased productivity and robustness, while reducing complexity. We have a proven record of impactful work in real-time and embedded programming models, with a special focus on real-time parallel programming, seamless programming of WSN, and new approaches for formal verification.

- modelling and analyzing temporal behaviour;
Today's platforms and technology requires guarantees in terms of performance, energy, and time. Systems must be as fast as possible, consume as less energy as they can, and deliver their functionality in a predictable way. To cope with these requirements, we address new problems resulting from the integration of distributed and many-core architectures in today’s platforms. We work on scheduling of parallel tasks in distributed, multi- and many-core platforms, sharing of resources (memories, communication networks, among others), use of dedicated computing accelerators such as GPGPUs, and task-to-processor mapping. Our outcomes also include probabilistic approaches, over classical analyses, to provide more efficient results.

- handling the requirements of mixed-criticalities;
Industry is facing new challenges with the integration of multiple applications with different levels of criticality within the same system. We are building upon our results for mixed-criticality systems and developing scheduling techniques to respect timing constraints of high criticality tasks and to recover from unexpected behaviours. We are also building analysis and design tools that enable a safe integration of applications of different criticalities in the same computing platform, eventually facilitating their certification.

- resource management in energy-aware computation;
Europe has a strong leadership in energy-aware embedded computing research with special focus in low-power platforms. With multi-core platforms becoming mainstream, there is a need to facilitate heterogeneous multi-cores which perform operations in a low-cost and efficient manner. On one hand, the challenge is how to balance energy and performance to provide for time-aware computing. On the other hand, this needs to be paired with mechanisms for temporal isolation. We addresses these challenges through accurate and predictable analysis models, which allow efficient management of system resources.

- real-time communication protocols that provide mobility, ubiquity, and pervasiveness;
The advent of advanced wireless technologies has enabled communication in previously difficult or uncharted environments. We have been leading international research in ubiquitous/networked embedded systems, with a special emphasis on Quality of Service (QoS) in low-power and low-cost wireless networks. The current focus is on designing efficient, reliable and real-time mechanisms under stringent application requirements and environmental conditions particularly, in the unlicensed and over-populated ISM band.

- combine “physical concerns” (like control systems, signal processing) and “computational concerns” (like complexity, schedulability, computability) with CPS Co-design;
Cyber-Physical Systems (CPS) integrate computation and physical processes, and are becoming pervasive in virtually every aspect of our daily life. Consequently, development of CPSs bring new and complex challenges in terms of design as these systems introduce safety and reliability requirements considerably different from those in general purpose computing and traditional real-time, embedded systems. We address the challenges posed by CPS in relevant fields such as scalable data aggregation, low power and embedded communication, and analysis of temporal behaviour of complex systems. Our researchers have adopted innovative co-design approaches to address scalable data aggregation in large-scale CPS, and are currently further exploring these concepts to propose novel distributed algorithms for CPS co-design.

- new demands at all layers of complex systems for better resource and QoS management.
The notion of Quality of Service (QoS) contracts and Service Level Agreements (SLA) is essential in order to handle the reliable and safe management of embedded distributed systems resources. Our focus is in the context of guaranteeing that the applications are able to get their desired resources as specified in SLA. Our results include architectures. mechanisms and analyses for scenarios that range from complex distributed systems (including middlewares for the Internet of Things) to resource constrained sensor nodes.