A discussion on channel sensing techniques By James Xu Supervised by Dr. Fakhrul Alam.

Channel Sensingfor Cognitive Radio

A discussion on channel sensing techniques

By James XuSupervised by Dr. Fakhrul Alam

Presentation overview

Introduction to wireless communications What is Cognitive Radio (CR)? Why do we need CR? What is channel sensing Our workbench setup Our research Our major contributions Conclusion

Wireless communications

What is Cognitive Radio (CR) Able to sense spectral environment Able to provide opportunistic access

Find gaps in the spectrumAdjust system parameters to utilize it

Why do we need CR?

Unlicensed spectrum is rare Almost none available under 3GHz

Why do we need CR? Licensed to primary user only (Incumbent) We are running out of space, and higher frequency

has problems We are under utilizing licensed space

CR is allowed usage (guarantee interference free)

What is channel sensing

The first step of CR is to identify free spectrum

Channel sensing is one of the most fundamental part of CR

What is channel sensing

Our research

Our research focus Energy detection Cyclostationary feature extraction

Our CR workbench

Energy detection

Easy to implement Fast Not effective under low SNR

What is energy detection? All radio transmission have energy

Based on hypothesis testing

Energy detector

Energy detector

Problems with the detector Misdetection on Narrowband signals

Problems with the detector Misdetection on partial signals

Adaptive energy detection We proposed an

improvement Introduced adaptive

detection

Improved detection

Adaptive ED characteristics

Adaptive ED characteristics

Cyclostationary Features

Another channel sensing strategy Digital communication systems have

built in periodicity A signal is a first order cyclostationary if

it’s mean is periodic A signal is second order cyclostationary

it it’s auto-correlation is periodic

Cyclostationary Features

Auto-correlation how much a signal has in common with itself, against delay

Alpha = Cyclic frequency S = Spectrum Correlation Function (SCF)

Cyclostationary Features

Work in Progress

This is preliminary, a proof of concept Could be done in future research

Our major contributions

Xu J. Y., Alam, F., Adaptive Energy Detection for Cognitive Radio: An Experimental Study, ICCIT, 2009

IET PATW Competition, Regional winner cr.jamesyxu.com Svn://cr.jamesyxu.com/svn CRLibs

CRLibs

A library for cognitive radio research

Conclusion

Workbench setup Investigated various channel sensing

techniques (ED, SCF, Spectral Entropy) Proposed and implemented

improvements (AED) Consolidated library (CRLibs), examples

and codebase (SVN) Full progress documentation (SVN, CR)

Acknowledgement

Dr Fakhrul Alam Dr James Chang Dr Tom Moir Everyone in our lab

Thank you for listening

You are invited to visit our CR workbench at Building 80

Much more details in the project report Adaptive energy detection model will be

on IEEEXplore end of December Any questions?