viswanath

Spectrum Sensing for Cognitive Radios

G Viswanath

Honeywell, Bangalore

•Overview and introduction to Cognitive Radios

•Approaches for spectrum sensing

•Conclusions

Outline

Motivation for Cognitive Radio

Spectrum Utilization

Ref: M.A.McHenry, “NSF Spectrum Occupancy Measurements Project Summary,” August

2005

Ghasemi and Sousa, IEEE Communications Magazine, April 2008

Increasing demand for spectrum

Existing scenario

– Under-utilization of spectrum

Innovative approach to improve spectrum

utilization

– Cognitive Radio

CR Scenario

• CR: Opportunistic Unlicensed Access

• To temporarily unused frequency bands (across the entire licensed radio spectrum)- A means to increase efficiency of spectrum usage

• Stringent safeguards required- On-going licensed operations should not be compromised

• Spectrum sensing based access- White spaces – primary user absent, and free of RF interferers

- Gray spaces – primary user absent but partially occupied by interferers

- Black spaces – primary user present

• Main functionality of Cognitive Radios- Ability to reliably identify unused frequency bands

- Sensing must be reliable and autonomous

• Radically different paradigm- Secondary (unlicensed) users - Opportunistic use of unused licensed bands

- Increased utilization of radio spectrum

TV Bands

Unlicensed Bands

• Several co-existing radios networks interfere with each other

- 2.4GHZ band

Zigbee, Bluetooth and wireless LAN

• Co-existence approaches critical for capacity

• The network geometry and its structural fluctuations are critical parameters that influence the performance of random networks.

• The interference and the signal strength at a receiver critically also depends on the distribution of the interfering transmitters

• Studying wireless networks based on geometry is discussed in http://users.ece.utexas.edu/~jandrews/stochgeom/index.htm

We look at signal processing approaches for spectrum sensing here

Spectrum Sensing

Methods of spectrum sensing

Aspects of Spectrum Sensing

Regulatory constraints

Spectrum Sensing Uncertainties

Spectrum Sensing Uncertainties

Detection Sensitivity

Detection Sensitivity

Spectrum Sensing - Approaches

Spectrum Sensing

Energy Detector

Performance of Energy Detector

Performance of Energy Detector (contd.)

Correlation Detector

Low Complexity Hybrid Detector for GSM

Key Question

• Can we look at a spectrum sensing algorithm that does not depend on:

- Estimate of noise variance

- Prior knowledge of the signal

Covariance Based Detector

Consider binary hypothesis testing problem again:

The transmitted signal is s(n) and the i.i.d white noise is

with variance )n(

2

Consider L consecutive samples of and define the following vectors

Covariance Based Detector

• Consider the statistical covariance matrices of the signal and noise:

•The off diagonal elements of covariance matrix are zero if the signal is not

present.

• If the signal is present and the signal samples are correlated then covariance

matrix is not a diagonal matrix.

•Consider the following:

Covariance Based Detector

• If there is no signal the ratio of T1 and T2 is ONE.

• If the signal is present the ratio is greater than one

• For a given probability of false alarm the threshold is set for the ratio T1 andT2

- The threshold is not related to noise power. Hence, robust to noise uncertainty

- The performance of the detector improves with the smoothing factor

- The performance of the detector also depends on number samples used for computing the sample autocorrelation

• Difficult to set the threshold based on probability of detection since signal is unknown

Spectral Covariance Based Sensing

• Spectral covariance based sensing (SCS)

- Exploits correlation of the signal and noise in frequency domain

- Test statistics computed from partial spectrogram and compared with a threshold

- 3dB performance improvement over covariance based detector for DTV signal detection

Key steps in the algorithm

• Down-convert the received signal to the base band

• Low pass filter and down-sample the received signal

• Compute the spectrogram of the signal

• Select the components near the DC terms for every dwell in the spectrogram

Key steps in the algorithm

• The reduced spectrogram matrix:

- Selects the spectral feature of the desired signal

- Reduces noise power

- Computational reduces

• Calculate sample covariance matrix

• Compute the test statistic:

A threshold is obtained for T1 and T2 based on the probability of false

alarm.

Improved SCS: Multi-band SCS System Model

Performance of multi-band SCS

Performance of MB-SCS algorithm for DTV signals

Performance with noise uncertainity

Key conclusions

• Spectrum sensing using energy detection requires estimate of noise uncertainity

• Correlation based techniques require signal model

• Spectrum sensing algorithm using statistical covariance

- Without signal knowledge

- Without estimate of noise uncertainty

• Next steps

- Co-operative spectrum sensing

- Cross layer based approaches for sensing

Thank You

Backup Slides

Software Defined Radios (SDR)

• Software Defined Radio (SDR)- A Software Defined Radio is a radio that is flexible

(programmable) to accommodate various physical layer formats and protocols

- A multiband, multimode radio with dynamic capability defined through software covering all layers of OSI protocol stack

Software Architecture

Reconfigurable

Generic Hardware

Flexible

Multiple Protocols

Upgradeable

Multiple Frequencies

Interoperable

Radio Architecture

Classification of SDRs

• Multi-band System

• Multi-standard System

• Multi-service system

• Multi-Channel System

SDR Drivers: NCW (Military) Vs ATM (Commercial)

• NCW in Military environment

• 33 Waveforms

• Key Initiative : JTRS program w/ clusters & incremental waveforms

- Cluster 1: Ground vehicles, Helicopters

- Cluster 2 : Hand-held

- Cluster 3 & 4 : Airborne, marine & fixed (AMF)

- Cluster 5 : Manpack/handheld radios

• ATM in commercial Airspace(NAS in U.S)

• 26 Waveforms across CNS

• Key Initiative : CNS/ATM Systems

- 3 radio cores (C,N & S) common across 3 segments AT

BRH

GA

- Or A single CNS radio Core across 3

segments?

Network Centric Warfare(NCW) Air Traffic Management (ATM)

NCW mirrors ATM; Priorities for Military & commercial differ!

Software Defined Radio to Cognitive Radio

• The FCC refers to a Software Defined Radio (SDR) as:

- “a transmitter in which the operating parameters … can be altered by making a change in software that controls the operation of the device without … changes in the hardware components that affect the radio frequency emissions.”

• The FCC view of cognitive radio:- “A cognitive radio (CR) is a radio that can change its transmitter parameters

based on interaction with the environment in which is operates.

Definition: Cognition

• According to Encyclopedia of Computer Science:

- Mental states and processes intervene between input stimuli and output responses

- The mental states and processes are described by algorithms

- The mental states and processes lend themselves to scientific investigations

Please note this is from a computational perspective

Why Cognitive Radios?

• Spectrum Utilization

- Presence of “Spectrum Holes”: band allocated to an user remains unused at a given time and geographical location

• Reliable communication

Definition: Cognitive Radio

• An intelligent reconfigurable radio that is aware of its surrounding environment

• Adapts its internal states to statistical variations in the incoming RF signal by making corresponding changes in the certain parameters to provide:

- Reliable communication

- Improved spectrum utilization

Cognitive Radio

• Responds to operators commands: “Turning the knobs”

• Also monitors its own performance continuously

Cognitive Radios: Tasks

• Reconfigurability: achieved through Software Defined Radio

• Other Cognitive tasks achieved using:

- Signal processing

- Machine learning

• Starts with passive sensing of RF stimuli and culminates with an action

Cognitive Radios: Tasks (contd.)

• Radio-scene analysis

- Estimation of “Interference Temperature”

Spectrum estimation techniques

- Detection of “Spectrum Holes”

Statistical techniques employed to detect

• Channel identification

- Estimation of channel state information

Blind and semi-blind approaches

- Prediction of channel capacity for use by the transmitter

• Co-operation

- Transmit power control and dynamic spectrum management

- Game theoretic approahes

• Dynamic Spectrum Sharing

Radio Scene Analysis

• Time-frequency analysis

• Multi-taper spectral estimation

- Optimal

• Large number of sensor to obtain the spatial variation

• Adaptive beamforming

- At the transmitter power is preserved

- At the receiver leads to improved interference cancellation

Transmit Power Control

• Optimal control theoretic based approach

• Game theoretic approach in the presence of competition

• Aims to:

- select the transmit power levels of n-unserviced users to jointly maximize their data-transmission rates, subject to the constraint of interference temperature

• Computationally feasible approach for a non-cooperative multi-user scenario:

- Maximize the performance of each unserviced transceiver subject to the constraint of interference temperature, irrespective what other transceivers do

Dynamic Spectrum Management

• Builds on the

- Spectrum holes detected

- Output of transmit power control

• Selects:

- A modulation strategy that adapts to the time varying conditions of the radio environment

• In OFDM case:

- Number of bits per channel varied based on the SNR

- Bandwidth and carrier frequency dynamically modified depending on “Spectrum Holes”

Illustration

A possible test bed

Consciousness in UWB networks

• Hybrid modeling for admission control

- A node leaves the network when there is no data to transmit, node failure or power exhaustion – discrete case

- Changes in the radio environment – Continuous case

Consciousness in UWB networks (contd.)

• In the uplink situation the time varying set of parameters are obtained based on

- Waveform used for pulse shaping

- Power level at base station

- Noise level at base station

- MUI

- Number of active nodes

Concluding Remarks

• All of the benefits of software defined radio

• Improved link performance

– Adapt away from bad channels

– Increase data rate on good channels

• Improved spectrum utilization

–Fill in unused spectrum

– Move away from over occupied spectrum

• New business propositions

–High speed internet in rural areas

–High data rate application networks (e.g., Video-conferencing)

• Significant interest from FCC, DoD

– Possible use in TV band reframing