self encoded spread spectrum

8/7/2019 Self Encoded Spread Spectrum

1/31

A N O V E L A P P R O A C H T O S P R E A D S P E C T R U M

C O M M U N I C A T I O N

Self encoded spread spectrum


2/31

Types of spread spectrum

y There are two common types of spread spectrum

Frequencyhopping spread spectrum (FHSS)

Direct sequence spread spectrum (DSSS)


3/31

FrequencyHopping Spread Spectrum

y Signal broadcast over seeminglyrandom series offrequencies

y Receiver hops between frequencies in sync with

transmittery Eavesdroppers hear unintelligible blips

y Jamming on one frequencyaffects onlya few bits


4/31

Types ofFHSS

y Frequencyshifted everyTc seconds

y Duration of signal element is Ts seconds

y SlowFHSS has Tc u Tsy Fast FHSS has Tc < Tsy Generallyfast FHSS gives improved performance in

noise (or jamming)


5/31

Direct Sequence Spread Spectrum (DSSS)

y Each bit represented bymultiple bits using spreading code

y Spreading code spreads signal across wider frequencyband In proportion to number of bits used

10 bit spreading code spreads signal across 10 times bandwidth of 1

bit codey One method:

Combine input with spreading code using XOR

Input bit 1 inverts spreading code bit

Input zero bit doesnt alter spreading code bit

Data rate equal to original spreading code


6/31

Direct Sequence Spread Spectrum Example


7/31

Walsh codes

y Walsh codes are used for synchronous DS-CDMA asspreading sequence so it is used in downlink channel

y One of a group of specialized PN codes having good

autocorrelation properties but poor cross-correlationproperties. Walsh codes are the backbone of theCDMAOne and cdma2000 cellular systems, and areused to support the individual channels usedsimultaneously within a cell. Walsh codes are

generated in firmware by applying the Hadamardtransform on 0 repeatedly.


8/31


9/31

Self encoded spread spectrum

y SESS is a novel wayto spread signal first publicationabout this scheme was done in September 1999

y In SESS present bit is spreaded using previous N Bits

as shown in following diagram:


10/31

Self encoded spread spectrumTransmitter

y Spreading code is obtained from source bits itself. Attransmitter delay registers are constantly updatedfrom an N-tap delay of data where Nis length of

spreading code.y Random nature of data from source is ensured by

applying proper source coding such e-g Lempel-Zivor Huffman codes, to remove redundancy in source

bits. Due to this factor bits of data stream can bemodeled as Bernoulli random variable.


11/31

Self encoded spread spectrumTransmitter

y Continued from previous slide

y As a result of applying source coding spreadingsequence is random and independent of current

symbol, but also dynamically changing from onesymbol to next.


12/31

Self encoded spread spectrumReceiver

y The operation of transmitter is reversed at thereceiver i-e recovered data is fed back to N-tap delayregister that provide estimated spreading code to de-

spread the signal at de-correlation detector.y Content of delay register of both transmitter and

receiver should be same at the start of communication.


13/31

Self encoded spread spectrumReceiver

y Now if the receiver is not synchronized withtransmitter or spreading is not known previouslythen it is really hard to detect the communication by

any unintended receiver making communicationreallysecure.


14/31

Analysis of SESS in AWGN environment

y The issue with SESS is that the performance isimpacted at low signal to noise ratios by errorpropagation. When the receiver detects a bit

incorrectly that error is inserted into the decodingsequence and it continues to affect the decodingprocess until it is shifted out of the sequence N bitslater.


15/31


y The signal attenuation depends on the chip length.This means for large N, a chip error would remain inthe register longer, but would contribute to a smaller

attenuation. Inversely, for a small N value a chiperror would rotate out of the register quickly, but

would contribute to a larger attenuation.

y Figure on next slide shows the effects of the sequence

length.


16/31


Figure 3: Effects ofChip Length on SESS


17/31


y Notice that the greater the sequence length the quickerthe SESS system converges to BPSK in an Additive WhiteGaussian Noise (AWGN) channel. Conversely, for lowerN values the BER stays closer to 0.5 longer.

y This makes sense when you consider that if there is anerror for a small N of 1 or 2, the next bit will bedemodulated with the wrong spreading code leaving witheither 100 or 50 percent of the wrong bits in the

demodulating spreading sequence. This leads to the 50percent error rate that can been seen in these two valueson Figure 3 on previous slide


18/31


y The chip errors in the receiver registers attenuate the de-spreaded signal strength. This can be regarded as a form ofself interference introduced by self encoding. Let X be therandom variable denoting the number of chip errors within

the receiver register of length N. For l chip errors, theamplitude attenuation can be expressed as:

y Then conditional probabilitycan be defined as:


19/31

Analysis of SESS inRayleigh fading environment

y The SESS system behaves similarly in a RayleighFading channel as the AWGN channel as seen inFigure 4 below.

Figure 4: Effects ofChip Length on SESS in

fading environment


20/31

Analysis of SESS inRayleigh fading environment

y O bserve in figure 4 how when N gets bigger, thesimulation converges toward the theoretical RayleighFading Channel line. This is the same behavior as is

see in the AWGN channel, in that, as N

theprobability of error approaches that of a BPSKsystem in that channel.


21/31

Iterative decoding

y Iterative decoding can be described as a decoding techniqueutilizing a soft-output decoding algorithm that is iteratedseveral times to improve the bit error performance of a codingscheme, with the goal of obtaining true maximum-likelihooddecoding, with less decoder complexity

y Because there is memory within the SESS modulation, it is anatural candidate for the Maximum Likelihood SequenceEstimation (MLSE) detection based on the Viterbi algorithm.MLSE detection improves the system performance byestimating the sequence of the received signals. However, thenumber of states in the Viterbi algorithm decoder grows

exponentially with the spreading factor, as can be seen in thetrellis diagram of SESS when N = 4 in Figure 5 on the nextSlide


22/31

Vterbi decoder for N=4

y There are 16 states for N=4 and theyincreaseexponentiallybythe factor of 2Nwhich isunacceptable complexityof the system.

Figure 5: Trellis Diagram of the Viterbi Algorithm for SESS of N=4


23/31

Iterative decoding

y An iterative detection scheme can be used instead toreduce the complexity to a linear order of thespreading factor, which achieves performance very

close to that of the MLSE detector.y The iterative detector is also able to be improved in

fading channels by adding a chip-interleaver asdiscussed in [1].


24/31

Iterative decoding procedure

y As describe in the previous section, the iterativedecoder has a complexity linear to that of thespreading code. The design used requires N+1

storage of the received data bits. The definition of aSESS systems states that the spreading codes aregenerated from the information being transmitted. If

we view the first bit after the encoder (called Bit 1),

then we can write the future N+1 bits (N is the lengthof the spreading code) as the part of the Bit 1 at thereceiver (see Figure 6 on the next slide).


25/31

Iterative decoding

1 =[ 10 , 11 , , 1 , 1(+1)]2 = [21 , 20 , , 2(+1), 2(+2)]3 = [32 ,31 , , 3(+2), 3(+3)]

= [

1

,

2

, ,

1

,

0

]

+1 = [+1 , +11 , , +1 2, +1 1]Fig 6. Decoding of signal

y From Figure 6 above, it is easyto see that 1 is not onlyrelated tothe previous N+1 bits, but also related to N future transmitted bits.

This means that N future bits contain information about 1 , thatcan be used to help make the final decision on Bit1 should there beexcessive channel noise or jamming on Bit1 that would normallycause an error.


26/31

Iterative decoding

y So by incorporating future transmitted signalstogether with previous detected bits, we expect toimprove the performance over the feedback detector,

which only estimates the current bits by correlatingwith N previous detected bits. How these futuretransmitted signals are incorporated into the finaldecision can be seen in Figure 7 on the next Slide.


27/31

Iterative decoding model

Figure 7 : Block diagram of iterative detection.


28/31

Performance ofIterative decoding

y The performance of the iterative detector can be seenin Figure 8.

Figure 8: Iterative Detector Performance in RayleighFading Channel


29/31

Performance ofIterative decoding

y From figure 8 it can be shown that system nearlyachieves the performance of BPSK in AWGNenvironment. For further improvement in the

performance diversity techniques like cooperativediversity can be introduced to reduce the effect ofdeep fades.


30/31

Reasons for using SESS

y As shown in previous figure previous (N-M) bitsspread the Nth bit, where M is the spreading factor.So spreading is stochastic in nature and it difficult to

de-spread the sequence byan unintended receiver.y This technique provide a feasible implementation of

random-coded spread spectrum systems thatpreviously have been thought to be impractical.


31/31

Conclusion

y From above we can conclude that SESS with iterativedecoding is promising solution for wirelesscommunication because it achieves low BERs in

fading environment and it is really secure form ofcommunication. It also has inherent time diversitybecause of using previous source bits for spreadingcurrent source bit.

self encoded spread spectrum

Documents