Gökhan Gül

M. Sc. Gökhan Gül

Technische Universität Darmstadt

Fachbereich Etit, Institut für Nachrichtentechnik

FG Signalverarbeitung, Raum S306/251

Merckstr.25, 64283 Darmstadt, Germany

FG Signalverarbeitung

+49 (0)6151 16 5571

+49 (0)6151 16 3778

Forschung:

DISTRIBUTED DETECTION IN COOPERATIVE SENSOR COMMUNICATIONS

Problem Description

Cooperative sensor communications is a new paradigm in the wireless communications area that allows geographically distributed nodes, terminals or sensors in a wireless network to share resources or to cooperate by means of distributed processing. One of the key roles of cooperative techniques is the ability to detect events of interest by means of distributed detection systems. This ability could serve as a sole purpose of the overall cooperative systems such as in wireless sensor networks for surveillance or could also serve as an initial goal such as in collaborative spectrum sensing for cognitive radio.

In parallel distributed detection with fusion center (FC) (as being one of the most popular distributed detection scheme), a set of dispersed nodes receive signals that contain information regarding a certain phenomenon which is perturbed by noise and sensing channel conditions such as fading and shadowing. Based on its observations, a node selects one of the possible messages according to a local decision rule and forwards it to the fusion center via a dedicated reporting channel which is also prone to various disturbance effects such as noise. Upon reception of the data from all designated nodes, the FC chooses a final decision based on a certain fusion rule.

The main focus of our research is:

  • Determination of optimum or sub-optimum decision and fusion rules for practical scenarios considering the conditions of the sensing and reporting channel.
  • Robust distributed detection in order to tackle the issue of the noise which deviates from the Gaussian assumptions in a distributed manner.
  • Correlated observations between nodes which could impose substantial computational complexity into the design of local decision and fusion rules.