As a follow-up from this seminar, a special issue of the Computer Communications journal from Elsevier was produced with selected contributions from the participants:
Special Issue of Computer Communications Wireless Sensor Networks and Applications - Proceedings of the Dagstuhl Seminar 04122.
Volume 28, Issue 13, Pages 1481-1602 (2 August 2005)
Editorial

1. Technical description

The advent of nano-technology has made it technologically feasible and economically viable to develop low-power devices that integrate general-purpose computing with multi-purpose sensing and wireless communications capabilities. It is expected that these small devices, referred to as sensor nodes, will be mass-produced and deployed, making their production cost negligible. Individual sensor nodes have a small, non-renewable power supply and, once deployed, must work unattended. For most applications we envision a massive deployment of sensor nodes, perhaps in the hundreds or even thousands. Aggregating sensor nodes into sophisticated computational and communication infrastructures, called sensor networks, will have a significant impact on a wide array of applications ranging from military, to scientific, to industrial, to health-care, to domestic, establishing ubiquitous wireless sensor networks that will pervade society redefining the way in which we live and work. Sensor networks are currently being established as a specific sub-task of the rapidly unfolding area of ubiquitous and pervasive computing.

The fundamental goal of a sensor network is to produce globally meaningful information from raw local data obtained by individual sensor nodes. Importantly, this goal must be achieved in the context of prolonging as much as possible the useful lifetime of the network and ensuring that the network remains highly available and continues to provide accurate information in the face of security attacks and hardware failure. The sheer number of sensors nodes in a sensor network, combined with the unique characteristics of their operating environment (anonymity of individual sensors, limited power budget and a possibly hostile environment), pose unique challenges to the design of sensor networks and their applications. For one thing, the limited power budget at the individual sensor node level mandates the design of ultra-lightweight communication protocols. Likewise, issues concerning how the data collected by individual sensor nodes could be queried and accessed and how concurrent sensing tasks could be executed internally are of particular significance. An important guideline in this direction is to perform as much local data processing at the sensor level as possible, avoiding the transmission of raw data through the sensor network. Indeed, it is known that it costs 3J of energy to transmit 1Kb of data a distance of 100 meters. Using the same amount of energy, a general-purpose processor with the modest specification of 100 million instructions/watt performs 300 million instructions. Recent advances in hardware technology are making it plain that the biggest challenge facing the sensor network community is the development of ultra-lightweight communication protocols ranging from training, to self-organization, to network maintenance, to security, to data collection and fusion, to routing, among many others.

There are several possible techniques that can be used to harvest the information produced by a sensor network. Perhaps the simplest involves using one or several sink nodes, special long-range radios, deployed alongside with the sensor nodes. In this scenario, the raw data collected by individual sensor nodes is fused, in stages, and forwarded to the sink nodes that provide the interface to the outside world. Sink nodes are also responsible for training the sensor network and for maintenance and repair operations. However, in some applications, it is impossible or impractical to deploy sink nodes within the sensor network. In such cases the task of harvesting the information produced by the sensor network and that of providing an interface to the outside world may be performed by aircraft and/or helicopters over-flying the sensor network, or by laser transmission to a satellite constellation. In this latter case, the bulk of the inter-sensor communications is by radio, since such communications are point to multi-point, while specialized sensors acting as local sinks communicate with the satellite constellation using laser beams.

While preserved energy can supply short-term applications, sensors dedicated to work over years may need to scavenge energy from the specific environment they are placed into, employing light, temperature, vibration, kinetics, magnetic fields, etc.

The ultra-lightweight protocols may leave not much room for advanced encryption schemes, so protection against overhearing in military applications and privacy protection in personal systems needs to be inherently built into the concepts from the beginning.

Reliability is expected to be a result of the large number of sensors deployed for a specific task. However, this can only be obtained if defective sensors can be excluded from the communication, and the sensors are calibrated -- either individually or collectively, either before deployment or continuously in their environment.

2. Goals of the seminar

The major stated goal of this Dagstuhl seminar is to bring together researchers from these groups and to provide a forum for a stimulating exchange of ideas and cross-fertilization between the participants. We welcome researchers from both industry and academia. Of a special interest to us is the participation of young researchers (35 years or younger). Schloss Dagstuhl is known to offer the ideal setting that will make this experience unique in all respects. We are convinced that the workshop-like relaxed atmosphere, typical of Dagstuhl seminars, will be conducive of high-caliber exchanges of ideas between the participants.

Topics of interest include but are not limited to:

• Energy-efficient training and self-organization
• Failure recovery and recalibration
• Resource management and Connection Admission Control (CAC)
• Media Access (MAC) protocols for wireless sensor network
• Network management scenarios and solutions
• Integration of sensor and terrestrial/satellite networks
• Energy-efficient routing
• Energy and complexity
• Environmental energy scavenging
• Thermodynamics
• Security issues in wireless sensor networks
• Performance evaluation
• Ubiquitous computing
• Applications

The discussions within the seminar are expected to address a large array of problems related to both theoretical and practical aspects of wireless sensor networks and their applications. In particular, presentation of practical implementations, experiments and ongoing projects is strongly encouraged.