DIGITAL SPREAD SPECTRUM SYSTEMS

1
DIGITAL SPREAD SPECTRUM SYSTEMS
ENG-737

• Wright State University
• James P. Stephens

2
FREQUENCY HOPPING

• Data is sent during the dwell time of a frequency
  hopping radio
• Modulation is typically Binary FSK
• The frequency shift is small compared to the
  frequency hop center frequency channels
• If the data is voice as in a tactical military
  radio or cordless telephone, it is digitized
  according to some digital voice standard
  (vocoder)
• Various vocoders have been adopted, but a common
  speech vocoder is known as CVSD (continuously
  variable, slope, delta) modulation
• Often, forward error correction (FEC) is
  employed, however, speech can tolerate
  considerable disruption before speech becomes
  unintelligible
• Speech data must be compressed to allow
  continuous transmission during time transmitter
  is transitioning to a new frequency

3
FREQUENCY HOPPINGExample

• CVSD speech ASICs often use 16 kbps, typically,
  for high quality speech
• If we wish to use employ frequency hopping, how
  much compression must we use?
• Assume the channel bandwidth (demodulator) can
  only support 20 kbps
• Then 16K/20K 0.80 ? 80 duty cycle
• If we need to send 100 bits per dwell, what is
  our hop rate?
• 100 bits (1/20K) 5 ms (Dwell time)
• 5 ms / 0.8 6.25 ms (Hop time) ? 160 hps

6.25 ms
100 data bits
5 ms
4
FREQUENCY HOPPINGClarifying Processing Gain

• A FH transmitter dwells for a period t1(time per
  hop) at each center frequency
• Hopping takes place over M frequencies
• PG Td BWss number of frequencies (M) ( for
  FH)
• Example
• Assume contiguous coverage, BWss 20 MHz
• N 1000 frequencies
• N 10 log 1000 30 dB
• If 20 MHz / 1000 20 kHz channel bandwidth
  (contiguous)
• PG 20 MHz / 20 KHz 1000 30 dB
• But not so if channels overlap or are
  non-contiguous

5
FREQUENCY HOPPER RECEIVER
st(t)
ht(t)
Sync is usually based on time-of-day and
correlation
1 . . . . .k
6
FREQUENCY HOPPER RECEIVER

• The frequency synthesizer output is a sequence of
  tones of duration Tc, therefore,
• ?
• ht(t) S 2p(t nTc) cos(?nt ?n )
• n - ?
• where p(t) is a unit amplitude pulse of duration
  Tc starting at time t 0
• ?nt and ?n are the radian frequency and phase
  during the nth frequency hop interval
• The frequency ?n is taken from a set of 2k
  frequencies

7
FREQUENCY HOPPER RECEIVER

• The transmitted signal is the data modulated
  carrier up-converted to a new frequency ( ?0 ?n
  ) for each FH chip
• ?
• st(t) sd(t) S 2p(t nTc) cos(?nt ?n )
• n - ?
• The transmitted power spectrum is the frequency
  convolution of Sd (f) and Ht (f)

8
FREQUENCY HOPPER RECEIVER

• Example
• FH, 250 hps, 2 ms dwell time, 48 bits per dwell
• Hop time 1 /250 4 ms
• ds 48 / 2 ms 24 kbps (signaling rate during a
  dwell)
• dr 48 / 4 ms 12 kbps (channel rate
  throughput)
• Minimum spacing for FSK tones are
• 1 / T 24 kHz (non-coherent FSK)
• 1 / 2T 48 kHz (coherent FSK)

9
FREQUENCY SYNTHESIZERS

• There are two fundamental techniques for
  implementing frequency synthesis
• Direct
• Indirect
• In the direct implementation, a number of
  frequencies are mixed together in various
  combinations to give all of the sum and
  difference frequencies
• Example
• cos(2??1) cos(2??2) 1/2 cos(2? (?1- ?2))
  1/2 cos(2? (?1 ?2))
• The selection is made based upon a digital
  control word as to which filters pass the
  selected tone
• The direct implementation becomes very difficult
  when a large number of frequencies must be used
• Size and weight of the filters are major factors
  in the choice to use this technique

10
SIMPLE DIRECT FREQUENCY SYNTHESIZER
11
BASIC ADD-AND-DIVIDE FREQUENCY SYNTHESIZER
A control word selects the gate on f2 fm which
are mixed with a reference frequency which
usually specifies the frequency separation or
spacing
12
INDIRECT SYNTHESIZERS

• Any synthesizer that employs a phase-locked loop
  is called an indirect synthesizer
• The output of the phase detector is filtered and
  drives a variable controlled oscillator (VCO)
• The phase detector drives the oscillator in the
  direction necessary to make ?? 0
• Any change causes the VCO to change in the
  opposite direction, thereby keeping the device
  locked to the input
• Frequency synthesis is accomplished by adding a
  divide-by-n block in the feedback path
• The VCO will lock to a multiple of the reference
  selected by n

13
BASIC INDIRECT FREQUENCY SYNTHESIZER
The divide-by-n is changed digitally by the code
generator to select another output frequency
14
NUMERICALLY CONTROLLED OSCILLATORS (NCO)

• More recent technique of frequency synthesizers
  are NCOs, also called direct digital
  synthesizers (DDS)
• DDSs are available as ASICs, see appendix 9 in
  text
• NCOs are available as FPGA cores, i.e. drop-in
  modules
• These devices simply have a sinusoid stored into
  memory that is outputted when selected.
• One such device uses a 32-bit tuning word to
  provide 0.0291 Hz tuning resolution and can
  change frequencies 23 million times per second,
  i.e.43 ns switching time
• These devices can control the phase, often with
  5-bits, in increments of 180, 90, 45, 22.5, 11.25
  degrees or combinations there of

15
BASIC NUMERICALLY CONTROLLED OSCILLATOR
16
DIRECT DIGITAL SYNTHESIZER
17
MULTIPLE CORRELATORS FOR FREQUENCY HOPPING
ACQUISITION
18
MULTIPLE CORRELATORS FOR FREQUENCY HOPPING
ACQUISITION
Time Delay
3 2 1 0
f1 f2 f3 f4 4
f4 f1 f2 f3 0
f3 f4 f1 f2 0
f2 f3 f4 f1 0
f1 f2 f3 f4 4
Delay
f1 f2 f3 f4
Let f1 101 MHz f2 107 MHz f3
105 MHz f4 103 MHz
Outcomes
19
REVISITING PROCESSING GAIN

• What is processing gain?
• From Peterson / Ziemer / Borth
• The amount of performance improvement that is
  achieved through the use of spread spectrum is
  defined as processing gain
• That effectively means that processing gain is
  the difference between a system using spread
  spectrum and system performance when not using
  spread spectrum. . .all else equal
• An approximation is
• Gp BWss / ri
• Some authors use other definitions
• Some system marketers use improper definitions
  just to make their system sound superior to
  competitors
• The particular definition chosen is of little
  consequence as long as it is understood that real
  system performance is the primary concern

20
REVISITING PROCESSING GAIN (Cont.)

• We could define processing gain as
• Gp td / tc
• Where td is the data bit time and tc is the chip
  time
• In the case of frequency hopping, a jammer or
  interferer can place all of his energy on a
  single narrowband signal, therefore, if the
  signal hops over M frequencies, the jammer must
  distribute power over all M frequencies with 1/M
  watts on each frequency
• Therefore, Gp M BWss / BWd (frequency
  hopping)
• however, we must assume contiguous,
  non-overlapping frequencies
• If overlapping occurs, Gp is reduced because the
  jammer can affect performance in adjacent
  channels. Thus Gp must be reduced by the amount
  of the overlap
• If non-contiguous, Gp gt M if jammer does not know
  system channelization since power will be wasted
  in regions where hopper never transmits

21
REVISITING PROCESSING GAIN (Cont.)

• Sklar defines processing gain as
• How much protection spreading can provide
  against interfering signal with finite power
• Spread spectrum distributes a relatively
  low-dimensional signal into a large-dimensional
  signal space
• The signal is thereby hidden so to speak in the
  signal space since the jammer does not know how
  to find it
• Dixon, however is not very consistent
• Page 6 A signal-to-noise advantage gained by
  modulation and demodulation process is called
  process gain
• Page 10 What is really meant by Gp in spread
  spectrum is actually jamming margin
• Gp BWss / BWinf (which assumes BWinf Rinf
  (1 Hz/bit))

22
REVISITING PROCESSING GAIN (Cont.)

• Note if
• Gp BWss / BWinf BWss / Rinf
• where Rinf 1 / Td
• Then Gp TdBWss (time-bandwidth product)

23
REVISITING PROCESSING GAIN (Cont.)

• Example
• Assume contiguous coverage for a frequency
  hopping radio
• BWss 20 MHz, N 1000 frequencies
• Gp N 10 log 1000 30 dB
• If
• 20x106 / 1000 20 kHz channelization
• Gp 20x106 / 20x103 1000 30 dB
• But not equivalent if channels overlap or are
  non-contiguous

24
COUNTERMEASURES
Electronic Attack (EA)

• To interfere with the enemys effective use of
  the electromagnetic spectrum
• Communications jamming involves the disruption of
  information, i.e. voice, video, digital
  command/control signals
• Rule One Jam receiver, not the transmitter

25
JAMMING MARGIN

• In general, the major factors which influence
  communicating in a jamming environment are
• Processing Gain
• Antenna gain (Tx, Rx, and jammer)
• Power (Tx and jammer)
• Receiver sensitivity and performance
• Geometrical channel
• Item 5 deals with issues such as directivity and
  line-of-sight features. Adaptive array
  processing and null steering are used to gain
  directivity advantages over a jammer or group of
  jammers

26
SIGNAL-TO-JAMMING RATIO

• Assume the jammer power dominates thermal noise
  (AWGN)
• The free-space propagation equation is
• (S/J)R PTGTGRdJ2 / PJGJdT2
• GR is the ratio of gain in the direction of the
  communication transmitter to gain in the jammer
  direction

27
SIGNAL-TO-JAMMING RATIO (Cont.)

• Since,
• (Eb/Jo) (S/J)R PG
• Where,
• (S/J)R the received signal energy-to-noise
  power spectral density ratio
• Then,
• (Eb/Jo) min required to achieve an acceptable
  PE performance must satisfy
• (Eb/Jo) min ? PTGTGR PG dJ2 / PJGJdT2
• Therefore, to improve performance we can increase
  PT, GT, GR, PG, or dJ
• Or decrease PJ, GJ, or dT

28
JAMMING STRATEGIES

• Noise
• Barrage
• Partial Band
• Narrowband
• Tone
• Single
• Multiple
• Swept
• Pulsed
• Smart
• Synchronized (coherent repeater)
• Non-synchronized (spectral matching)
• Knowledge based

29
PROBABILITY OF BER VERSUS SNR
Digital signals are highly susceptible to gradual
degradation
BER
SNR (Eb/N0)
30
KNOWLEDGE POWER RELATIONSHIP IN JAMMING
Brute Force Jamming
Power Required to Jam Victim
Smart / Responsive Jamming
Knowledge Required About Victim
31
JAMMING TECHNIQUES
32
JAMMING TECHNIQUES (Cont)
33
JAMMING TECHNIQUES (Cont)
G3
G2
G1
WSS
STEPPED TONES
34
DSSS IMMUNITY TO WIDEBAND NOISE
Noise jammer rejected by receiver

• Least power efficient technique but more covert
  than CW
• Requires no knowledge of signal
• High collateral damage (fratricide)
• Jamming power may be adjusted for gradual
  degradation

35
DSSS Performance in Broadband Noise Jamming

For BPSK modulation
where
For NoJo
J/S jamming/signal ratio Gp processing
gain
36
DSSS Performance in Broadband Noise Jamming
37
DSSS IMMUNITY TO CW
CW Interferer rejected by receiver

• Requires high power to overcome DSSS processing
  gain
• More power efficient than wideband noise
• Non-covert, target may employ filter to remove
  jammer

38
DSSS Performance in Tone Jamming

N Processing gain S signal power
Pt noise power Tb
data bit duration phase angle difference
between jammer and target signal
frequency difference between jammer and target
signal Pj power of jammer tone
39
DSSS Performance in Tone Jamming
40
DSSS Performance in Pulse Jamming
for 0ltplt1
Po for optimal Pe
if Jo gtgtNo
and 1
and
41
DSSS Performance in Pulse Jamming
42
JAMMING STRATEGIES AGAINST DSSS

• Most effective (non-adaptive) technique is
  provided by single-tone jammer at or near the
  carrier frequency
• This stresses the carrier suppression of balanced
  demodulators
• CCM
• Use an adaptive notch filter to delete the tone
• Detect the tone by a PLL and then subtract it
  from the signal or spatially null the jammer
• Decipher the PN code, replicate it as a jamming
  signal which will not be eliminated by the
  processing gain
• Most effective if jammer can become synchronized
  to the receiver
• CCM
• Make the PN code generators programmable so that
  the code can be readily changed or use complex,
  adaptive, codes

43
JAMMING STRATEGIES AGAINST DSSS (Cont.)

• Determine the carrier frequency and chip rate,
  then jam with a PN signal having these parameters
  (spectral matching)
• Less effective than 1) or 2), but more difficult
  to counter
• CCM - Use an adaptive code rates (ditter)
• Attack the acquisition process using a
  combination of 1) or 3)
• CCM Use short code for quick acquisition, then
  switch to longer code
• Pulse jamming and swept jamming at the carrier
  frequency
• Generally less effective than other methods
• Can be vary effective against AGC and tracking
  loops of target receiver if knowledge of receiver
  design is known
• CCM Use interleaving and error corrective
  coding

44
JAMMING STRATEGIES AGAINST FH

• Repeater jamming which involves intercepting
  signal, determining the center frequency, and
  transmitting a tone at that carrier frequency
• Very effective against slower FH systems
• CCM
• Increase hop rate
• Partial band or multitone
• Jammer places a series of tones across bandwidth
  where the received power per jamming tone exceeds
  the systems received power per hop
• CCM
• Use error corrective coding with interleaving
• Swept frequency
• Increases the BER, but is less effective than 1)
  or 2)
• CCM
• Use error corrective coding with interleaving
• Note Generally speaking, FH systems are less
  susceptible to attacks on acquisition than are
  DSSS

45
THE TACTICAL SCENARIO
Hopper Link
Jamming Link
Monitor Link
46
GEOMETRY FOR FREQUENCY HOP REPEAT JAMMER

• Th is the hopping period and ? is the fraction
  of hopping period within which the jammer must
  operate to be effective (Typically 50 of the
  dwell time)

47
GEOMETRY FOR FREQUENCY HOP REPEAT JAMMER

• For jamming to be effective we must have
• d2 d3 d1

Propagation time for Jammer
Where, tp jammer processing time c speed
of light (3 x 108 m/sec) (1 - ?) fraction of
dwell to be jammed
Source Modern Communications Jamming Principles
and Techniques - Poisel
48
HOP RATE VERSUS STAND-OFF DISTANCE