universität siegen institut für digitale kommunikationssysteme univ.-prof. dr. christoph ruland...

Universität SiegenUniversität SiegenInstitut für Digitale KommunikationssystemeInstitut für Digitale KommunikationssystemeUniv.-Prof. Dr. Christoph RulandUniv.-Prof. Dr. Christoph RulandHölderlinstraße 3Hölderlinstraße 3D-57076 SiegenD-57076 Siegenhttp://www.dcs.uni-siegen.dehttp://www.dcs.uni-siegen.de

FEC over Wireless Channels using FEC over Wireless Channels using Cryptographic Check Values as Coding Cryptographic Check Values as Coding

RedundancyRedundancy

Belgrade, April 2, 2009Belgrade, April 2, 2009Nataša ŽivićNataša Živić

Wireless Communication SystemsWireless Communication Systems

Wireless communication systems have found heavy Wireless communication systems have found heavy deployments over recent yearsdeployments over recent years

Many popular applications have emerged for wireless Many popular applications have emerged for wireless networks including, voice calls, SMS, MMS, web browsing networks including, voice calls, SMS, MMS, web browsing and video conferencing, to name a few.and video conferencing, to name a few.

These services not only needs to be provided but rather These services not only needs to be provided but rather provided in a secure and reliable mannerprovided in a secure and reliable manner

Our work focus on reliablity of data transfer, from source to Our work focus on reliablity of data transfer, from source to destination, using security mechanisms as a tooldestination, using security mechanisms as a tool

18.04.2008 ITG-Fachgruppensitzung 2

Channel Coding & CryptographyChannel Coding & Cryptography

We consider the following two basic limitations of a wireless We consider the following two basic limitations of a wireless channel,channel,

1.1. Wireless channel is inherently prone to transmission errorsWireless channel is inherently prone to transmission errors Forward Error Correction (FEC) is used as a solution in those Forward Error Correction (FEC) is used as a solution in those

scenarios where the feedback link is missing and thus Automatic scenarios where the feedback link is missing and thus Automatic Repeat Request (ARQ) can not be usedRepeat Request (ARQ) can not be used

2.2. Wireless channel is prone to the problem of data security Wireless channel is prone to the problem of data security tootoo Cryptography and digital signatures are used as a solution to these Cryptography and digital signatures are used as a solution to these

problemsproblems Cryptography solves the problems of eves-dropping & digital Cryptography solves the problems of eves-dropping & digital

signatures is a solution to repudiation by the sendersignatures is a solution to repudiation by the sender

18.04.2008 ITG-Fachgruppensitzung 3

Joint Channel Coding & CryptographyJoint Channel Coding & Cryptography

So far, data security and the correction of errors induced by So far, data security and the correction of errors induced by the transmission channel have been dealt separatelythe transmission channel have been dealt separately

In this work we present a novel technique which combines In this work we present a novel technique which combines these two different areas together to give a joint solution to these two different areas together to give a joint solution to both the problems discussed beforeboth the problems discussed before

We use cryptography (in the form of cryptographic check We use cryptography (in the form of cryptographic check functions) to improve the BER over the channelfunctions) to improve the BER over the channel

Moreover we use the channel coding to improve the results Moreover we use the channel coding to improve the results of cryptography as wellof cryptography as well

We call this technique as Joint Channel Coding & We call this technique as Joint Channel Coding & CryptographyCryptography

18.04.2008 ITG-Fachgruppensitzung 4

SOURCESOURCE

ENCODERCHANNELENCODER

MODULATOR

DEMODULATORCHANNELDECODER

SINK

Data Communication (Building Blocks)

SOURCEDECODER

L-values

NOISY CHANNEL

Data (Text, Audio, video)

Received Data (probably with errors)

Objective

MODULATORCHANNELENCODER

SOURCEENCODER

SOURCE ENCRYPTOR

DECRYPTOR DEMODULATORCHANNEL DECODER

SOURCEDECODER

SINK

L-values

L-values

Data (Text, Audio, video)

Received Data (probably with errors)

NOISY CHANNEL

CHANNEL DECODER(SISO)

SOFT INPUT DECRYPTOR

SID block L (1.bit) L (2.bit) L (3.bit)…

Soft Input

L = L-Values(Obtained from probability distribution)

Soft Input DecryptionSoft Input Decryption

Soft Input Decryption AlgorithmSoft Input Decryption Algorithm

Channel

SISO Channel Decoder

Number of trials exceeded

the max?

No

Source Decoder

No

Bit flipping

No Successfully verified data

Verification successful?

Unsuccessfully verified data

(contains errors) Yes

L Values of bits

Ordering of bits into an increasing array of their |L|-

values

Decryptor (Verification)

Soft Input Decryption Process

Feedback information about correct decoded bits

CHANNEL DECODER(SISO)

SOFT INPUT DECRYPTOR

SID block L (1.bit) L (2.bit) L (3.bit)…

Soft Input

Soft Input Decryption with FeedbackSoft Input Decryption with Feedback

ma || na

block a

mb || nb

block b

a b … b a b … b a … … a b … b

u

…V: …

ma mb

na nb

CCF CCFkey key

CCF – Cryptographic Check Function

Assembling message u for TransmissionAssembling message u for Transmission

Parallel Soft Input Decryption AlgorithmParallel Soft Input Decryption Algorithm

Demodulation

Channel

SID of Block a‘ SID of Block b‘

SISO Channel Decoding

Block a‘ Block b‘

BER1,SID BER1,SID

SISO Channel Decoding

BERfeedback

SID of Block a‘ SID of Block b‘

BER2,SID

Step

1

Step 1

Step 2

Step 3BER2,SID

BERcd1

SISO Channel Decoding

Segmentation of u‘

(into blocks a‘ & b‘)

BERfeedback

Encoder: - Convolutional encoder (1/2) BPSK modulation AWGN channel Decoder: - Convolutional MAP decoder (1/2 and 1/3) Length of SID Blocks: 192 bits (128 bit message and 64 bit cryptographic

check value) 8 (16) bits with the minimal absolute L-values are tested Eb/N0 increased in steps of 0,5 dB For each result (point of curves) 50 000 tests are done

Simulation ParametersSimulation Parameters

Simulation ResultsSimulation Results

Joint Soft Input Decryption and Source Channel Coding

Improvement of arithmetic efficiency of Soft Input Decryption method

Development of new strategies for Soft Input Decryption

Extension of Soft Input Decryption with Feedback to Turbo principle

Analysis of the impact of the channel encoder to Soft Input Decryption

SISO decryption

SummarySummary

Thank you very much

for your attention!