Cooperating Embedded Systems, 7.5 credits

Introduction

Motivation

We are today flooded by many small and cheap computing devices that are possible to use also in a mobile setting. Ubiquitous computing is a term that refers to the use of embedded processing in everyday objects. In the ubiquitous computing vision objects often are mobile and communicate with each other and the user by means of ad hoc wireless networking. Pervasive computing is a similar term that refers to the huge amount of cheap devices. Nomadic computing refers to the way people using them are mowing around and use such devices in different areas and contexts in a semi-mobile way.

Ambient intelligence is another popular term that refers to an environment that also is sensitive, adaptive, and responsive to the presence of people. Such ambient intelligence or awareness within ubiquitous, pervasive, mobile and embedded computing systems available almost everywhere enables new or better applications and services as well as performance improvements. Sentient computing is a similar concept used in conjunction with smart spaces and context-aware smart objects supporting us everywhere. From a control initiative perspective we can identify client-server, master-slave and peer-peer computing structures sometimes with brokers and mediators as glue. In a not so far future one can envision self-organizing amorphous nano-scale processing structures. Organizing the cooperation between the nodes in these kinds of networks creates new requirements on programming and system technologies that can glue all these things together.

This area has during the last years been given much attention inspired by different visions and plans regarding new or enhanced, often position, location and/or context aware, self-configuring and self-healing applications based on collaborative signal and information processing in wired and wireless networks of computing sensor and actuator nodes embedded or placed - almost everywhere - in the environment. Being embedded systems, the proposed solutions also must be able to deal with many different types of design constraint's. To enable some of the visions autonomous and decentralized forms of distributed system solutions will be needed. Cooperation among more or less autonomous nodes to solve a common goal is one possibility. Inspiration can be found in the principles for both cooperative and non-cooperative games.

There are many viable system architecture solutions. The nodes in a network of cooperating embedded systems may for example contain a combination of sensor, actuator, processing and communication devices or can be separated in specialized sensor devices in a sensor network, control devices in a control network and actuator devices in an actuator network. Connections between nodes and networks are sometimes wired in predefined structures and in other cases wireless and more self organizing.

A network of cooperating embedded systems can for example be used as an active multi-element information collector where cooperation is used for synchronization, fusion and refinement of the sensed and collected information. Much of this kind of computing highlight's the need for energy efficient stream based processing in tightly coupled processor systems. Sensor networks can also be directly attached or embedded with objects or goods for localization, tracking and surveillance of these items, in the real geographical space or in a logical supply chain. Other networks could be there to supervise or help elderly persons in their daily life or help to secure our office or home environments. Another important area is climate and environment monitoring systems. If one assumes that these networks are built on the use of many cheap embedded system nodes one can then also envision that several traditional problems need to or can be solved in new ways.

Goal

More specifically the aim of this course is to understand what the research issues are and what new requirements that wired and wireless sensor and actuator networks (as well as concepts and visions such as amorphous, ambient, ubiquitous and pervasive computing) give rise to. Some of these issues are: energy supply, scavenging and saving; availability, fault tolerance and robustness; clock synchronization and real-time behaviour; coordination of operations on data and other resources; context awareness, localization and positioning; data filtering and data fusion; efficient processing and programming of stream applications; data consistency, security and integrity; software components and middleware; partitioning in central and decentralized functions; media access protocols; collaborative multi-path and multi-hop message passing; as well as choices between ad hoc vs. planned networks structures and routing.

The course gives a wide - architecture level perspective - on a set of research issues that are interrelated by the term cooperating embedded systems.