Embedded Systems - Introduction
by Erwin Schoitsch
Each day, our lives become more dependent on 'embedded systems', digital information technology that is embedded in our environment. This includes not only safety-critical applications such as automotive devices and controls, railways, aircraft, aerospace and medical devices, but also communications, 'mobile worlds' and 'e-worlds', the 'smart' home, clothes, factories etc. All of these have wide-ranging impacts on society, including security, privacy and modes of working and living. More than 98% of processors applied today are in embedded systems, and are no longer visible to the customer as 'computers' in the ordinary sense. New processors and methods of processing, sensors, actuators, communications and infrastructures are 'enablers' for this very pervasive computing. They are in a sense ubiquitous, that is, almost invisible to the user and almost omnipresent. As such, they form the basis for a significant economic push.
These applications are 'vision driven', as in the following examples:
- Automotive: Accident free driving
- Aerospace: A free, safe sky for all
- Medical Devices: Robotic surgeon
- Communications: Seamless connectivity
- e-Life: ubiquitous/pervasive computing.
The European Context
The European Commission has recognised the importance of embedded systems by creating a new unit in the IST Directorate. The visions surrounding the AMI-space (embedded systems everywhere, described in the context of human life as 'ambient intelligence') have considerably influenced the 6th Framework Programme of the IST domain. In this issue we focus on hard real-time, dependability/safety and AMI-scenario applications. There also exists a separate strategic objective on embedded systems in the work program 2003-2004, namely, to develop the next generation of technologies and tools for modelling, design, implementation and operation of hardware/software systems embedded in intelligent devices. An end-to-end systems vision should allow cost-efficient systems to be built with optimal performance, high confidence, reduced time to market and faster deployment. The focus is on the following:
- Middleware and platforms for building networked embedded systems that aim to hide the complexity of underlying computing, communications, sensing and control, while at the same time providing efficient and effective distribution of resources at low cost. The emphasis will be placed on middleware for small wireless devices, eg mobile phones or PDAs, which make the design, programming, verification and maintenance of systems including such devices easier. Effort will also be devoted to scalable and self-organising platforms that offer services for ad-hoc networking of very small devices and for mastering complexity through perception techniques for object and event recognition and advanced computing and control.
- Concepts, methods and tools for system design and development of warrantable software components and implementation of systems, with an emphasis on the correct handling of complex real-time constraints. Work includes unification of computational models and composition methods, holistic design addressing event and time constraints, interface technologies in hard- and software addressing real-world and legacy issues, and techniques and integrated validation tools to ensure ultra-stable, dependable embedded systems.
- Advanced controls for real-time systems with an emphasis on hybrid systems theories including non-linear processes with both constraints and switching modes. Advanced controls for networked embedded systems, focussing on networked autonomous and fault-adaptive control and management, as well as on reasoning, behaviour, global performance and robustness.
This strategic objective is covered in the second call (15 June to 15 October 2003). The recently founded ERCIM Working Group 'Dependable Software-Intensive, Embedded Systems' is involved in a large IP proposal (Integrated Project, one of the new instruments of European funded research) called DECOS, Dependable Embedded Components and Systems (see article on page 22).
New Research Challenges
A new set of research challenges is presented due to the enormous dimensions of deployment and connectivity, which imply new levels of complexity and risk and new failure modes:
- context-awareness (identify, localise and interact with persons and objects, in a non-location-dependent manner)
- intelligence (the digital environment adapts to mobile objects and persons, and learns and interacts independently, thus providing fascinating new services)
- natural interaction (human language, gestures, speech synthesis)
- personalisation (user-centred, dynamic adaptation to changing situations and user profiles/preferences)
- dependability (time dynamics, timely responsiveness, security, safety, availability, maintainability, robustness etc)
- restricted resources (low power management, size, weight, memory, processing power/speed, interfaces etc)
- wireless/mobile, seamless communication
- hard real-time applications (automotive, aerospace, railways, process control, medical devices, communications etc)
- challenges of composability, COTS, (ultra-)high confidence in design and certification, security and predictability.
Important sub-areas are (rigorous) hardware/software co-design, smart new sensors/actuators, continuous connectivity issues and limited resource management. Horizontal issues include dependability, system integration, software technologies, standardisation, certification, and critical infrastructures.
About this issue
This issue of ERCIM News covers all aspects of embedded systems, technical as well as broadly socio-economical, and gives an overview of the scientific and engineering challenges, and fascinating possibilities and risks offered by this enabling technology.
The contributions are grouped into:
- Applications (6): Railways, Remote Control of Scientific Experiments, a more general article on 'Emerging Frontiers in Embedded Systems', Wearable Systems for Everyday Use, Embedded Systems in Cars, and Application in Transport Logistics
- Networking and EU-Projects (2)
- System Architectures, Methods, Languages and Tools (11)
- Education (3).
These contributions clearly demonstrate that the scientific community is attracted primarily by the interesting system, design and implementation challenges of embedded systems. It is also becoming evident that our educational systems require new curricula, labs and courses, as well as networks for experiential exchange, in order to keep up with such innovation. The section on applications shows that the driving force behind this comes from critical control applications (automotives, railways etc) and the 'everyday ubiquitous computing (AMI)' community.
ARC Seibersdorf Research (AARIT)
Tel: +43 50550 3117