Lecture: Hot Topics in Communications Systems, Sensor-Networks, and Pervasive Computing

Other:

• Term: Master Computer Science, Diploma (Graduate)
• Requirements:
• Faculty:
• Effort:

10.06.2008 17:00-19:00, Hörsaal I

Sebastian Max (RWTH Aachen)
"Radio Network Planning for Wireless Mesh Networks"

Radio Network Planning (RNP) is an essential element of 2/3G radio
network deployment: Before (and also during) network operation, the
provider uses an RNP tool to optimize the network configuration. This
optimization works in three steps: (i) the essential properties of the
network parametrize a predefined model, (ii) performance metrics are
derived by the model and (iii) the configuration is changed. By using
the performance metrics as a utility function, an iterative
optimization becomes possible.

In contrast to this approach, IEEE 802.11-based Wireless Mesh Networks
(WMNs) are usually deployed using a trial-and-error strategy: Access
Points are positioned for coverage without taking interference into
account; performance is measured after the network setup for the first
time; etc.

A major challenge towards RNP for WMNs is step (ii) of the optimization
process: For this purpose, the characteristics of the IEEE 802.11 Medium
Access Control (MAC) in WMNs have to be analysed and modelled. In the
talk, a MAC model tailored specifically for this purpose is
introduced and evaluated.

26.06.2008 17:00-19:00, Hörsaal II

Jörg Hähner (Uni. Hannover)
"Self-Organizing Distributed Smart Camera Systems"

Future surveillance systems are expected to rely on distributed vision
networks that consist of Smart Cameras (SCs). Each SC is an autonomous
node containing a PTZ (pan/tilt/zoom) camera, processing capabilities
(CPU, memory, etc.), and a communication interface. The talk aims at
describing the design and evaluation of protocols and algorithms that
make way for self-organisation in distributed vision networks. The term
self-organisation implies self-configuration (SCs calibrate and adjust
their fields of view by their PTZ function to search/track objects),
self-optimisation (SCs adjust their fields of view to reach high
surveillance coverage), self-healing (SCs take over tasks from failing
SCs), self-protection (functional monitoring) and anticipation (SCs
detect critical events and raise alarms). We consider SCs to form
distributed systems that rely on wireless mesh networks. This
architecture allows to overcome drawbacks in terms of scalability and
fault tolerance of centralised vision networks as used today.

08.07.2008 17:00-19:00, Hörsaal I

Carsten Buschmann (Uni. Lübeck)
"Funkbasierte Distanzschätzung für Sensornetzwerke und ihre Anwendung zur Lokationsbestimmung in Sensornetzen"

Die Position, an der Ereignisse detektiert wurden, ist für viele
Sensornetzanwendungen eine wichtige Kontextgröße. Um Sensordaten
entsprechend verorten zu können, müssen die Sensorknoten folglich ihre
Position bestimmen. Als entsprechendes Verfahren hat sich die
Multilateration durchgesetzt, die als einen wesentlichen Schritt die
Distanzschätzung umfasst. Der Vortrag stellt Verfahren zur
Distanzschätzung sowie zur Positionsbestimmung vor, berichtet von
praktischen Erfahrungen mit der Distanzschätzung, sowie entsprechenden
Modellbildungen und Simulationen.