This action might not be possible to undo. Are you sure you want to continue?

**Shiv Kalyanaraman Google: “Shiv RPI” shivkuma@ecse.rpi.edu
**

Based upon slides of Sorour Falahati, A. Goldsmith, & textbooks by Bernard Sklar & A. Goldsmith

Rensselaer Polytechnic Institute

Shivkumar Kalyanaraman

: “shiv rpi”

1

The Basics

Rensselaer Polytechnic Institute

Shivkumar Kalyanaraman 2

: “shiv rpi”

Big Picture: Detection under AWGN

Rensselaer Polytechnic Institute

Shivkumar Kalyanaraman 3

: “shiv rpi”

Additive White Gaussian Noise (AWGN) Thermal noise is described by a zero-mean Gaussian random process. hence. n(t) that ADDS on to the signal => “additive” Its PSD is flat. it is called white noise. Autocorrelation is a spike at 0: uncorrelated at any non-zero lag [w/Hz] Power spectral Density (flat => “white”) Rensselaer Polytechnic Institute Probability density function (gaussian) Autocorrelation Function (uncorrelated) Shivkumar Kalyanaraman 4 : “shiv rpi” .

with a random noise vector w Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 5 : “shiv rpi” .Effect of Noise in Signal Space The cloud falls off exponentially (gaussian). Vector viewpoint can be used in signal space.

Maximum Likelihood (ML) Detection: Scalar Case “likelihoods” Assuming both symbols equally likely: uA is chosen if Log-Likelihood => Rensselaer Polytechnic Institute A simple distance criterion! Shivkumar Kalyanaraman 6 : “shiv rpi” .

AWGN Detection for Binary PAM pz (z | m2 ) pz (z | m1 ) s2 − Eb 0 s1 Eb ψ 1 (t ) ⎛ s1 − s 2 / 2 ⎞ ⎟ Pe (m1 ) = Pe (m2 ) = Q⎜ ⎜ N /2 ⎟ 0 ⎝ ⎠ ⎛ 2 Eb PB = PE (2) = Q⎜ ⎜ N 0 ⎝ Rensselaer Polytechnic Institute ⎞ ⎟ ⎟ ⎠ Shivkumar Kalyanaraman : “shiv rpi” 7 .

Bigger Picture General structure of a communication systems Noise Info. signal signal Receiver Transmitter Channel User Transmitter Formatter Source encoder Channel encoder Modulator Receiver Formatter Rensselaer Polytechnic Institute Source decoder 8 Channel decoder Demodulator Shivkumar Kalyanaraman : “shiv rpi” . SOURCE Source Received Received Transmitted info.

Digital vs Analog Comm: Basics Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 9 : “shiv rpi” .

Voice Data Media Rensselaer Polytechnic Institute A bit is a bit! Shivkumar Kalyanaraman 10 : “shiv rpi” .Digital versus analog Advantages of digital communications: Regenerator receiver Original pulse Regenerated pulse Propagation distance Different kinds of digital signal are treated identically.

but also are amplified or attenuated equally. Output Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 11 : “shiv rpi” .Signal transmission through linear systems Input Linear system Deterministic signals: Random signals: Ideal distortion less transmission: All the frequency components of the signal not only arrive with an identical time delay.

Signal transmission (cont’d) Ideal filters: Low-pass Non-causal! Band-pass High-pass Duality => similar problems occur w/ rectangular pulses in time domain. Realizable filters: RC filters Butterworth filter Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 12 : “shiv rpi” .

Bandwidth of signal Baseband versus bandpass: Baseband signal Local oscillator Bandpass signal Bandwidth dilemma: Bandlimited signals are not realizable! Realizable signals have infinite bandwidth! Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 13 : “shiv rpi” .

Bandwidth of signal: Approximations Different definition of bandwidth: a) Half-power bandwidth b) Noise equivalent bandwidth c) Null-to-null bandwidth d) Fractional power containment bandwidth e) Bounded power spectral density f) Absolute bandwidth (a) (b) (c) (d) Rensselaer Polytechnic Institute Shivkumar Kalyanaraman (e)50dB 14 : “shiv rpi” .

Format Sample Quantize Encode Pulse modulate Pulse waveforms Transmit Bit stream Format Analog info. Low-pass filter Decode Channel sink Textual info. Analog info. Textual source info. Digital info. Demodulate/ Detect Receive Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 15 : “shiv rpi” .Formatting and transmission of baseband signal Digital info.

Sampling of Analog Signals Time domain Frequency domain xs (t ) = xδ (t ) × x(t ) x(t ) X s ( f ) = Xδ ( f ) ∗ X ( f ) | X(f )| xδ (t ) | Xδ ( f ) | xs (t ) | Xs( f )| Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 16 : “shiv rpi” .

Aliasing effect & Nyquist Rate LP filter Nyquist rate aliasing Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 17 : “shiv rpi” .

Undersampling &Aliasing in Time Domain Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 18 : “shiv rpi” .

Key: use sinc() pulses for reconstruction/interpolation Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 19 : “shiv rpi” .Nyquist Sampling & Reconstruction: Time Domain Note: correct reconstruction does not draw straight lines between samples.

Nyquist Reconstruction: Frequency Domain The impulse response of the reconstruction filter has a classic 'sin(x)/x shape. The stimulus fed to this filter is the series of discrete impulses which are the samples. Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 20 : “shiv rpi” .

Sampling theorem Analog signal Sampling process Pulse amplitude modulated (PAM) signal Sampling theorem: A bandlimited signal with no spectral components beyond . In practice need to sample faster than this because the receiving filter will not be sharp. Shivkumar Kalyanaraman 21 : “shiv rpi” Rensselaer Polytechnic Institute . is called Nyquist rate. can be uniquely determined by values sampled at uniform intervals of The sampling rate.

Out In Average quantization noise power Quantized values Signal peak power Signal power to average quantization noise power Shivkumar Kalyanaraman 22 : “shiv rpi” Rensselaer Polytechnic Institute .Quantization Amplitude quantizing: Mapping samples of a continuous amplitude waveform to a finite set of amplitudes.

Pulse code modulation (PCM): Encoding the quantized signals into a digital word (PCM word or codeword).Encoding (PCM) A uniform linear quantizer is called Pulse Code Modulation (PCM). Each quantized sample is digitally encoded into an l bits codeword where L in the number of quantization levels and Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 23 : “shiv rpi” .

Quantization error Quantizing error: The difference between the input and output of a quantizer Process of quantizing noise Qauntizer ˆ e(t ) = x(t ) − x(t ) y = q( x) AGC Model of quantizing noise x(t ) x ˆ x(t ) x(t ) e(t ) ˆ x(t ) + e(t ) = ˆ x(t ) − x(t ) Shivkumar Kalyanaraman : “shiv rpi” Rensselaer Polytechnic Institute 24 .

compression+expansion companding y = C ( x) x(t ) ˆ x y (t ) x ˆ y (t ) ˆ y Qauntize Expand ˆ x(t ) Compress Rensselaer Polytechnic Institute Transmitter 25 Channel Shivkumar Kalyanaraman : “shiv rpi” Receiver . At the receiver. an inverse compression characteristic.Non-uniform quantization It is done by uniformly quantizing the “compressed” signal. called “expansion” is employed to avoid signal distortion.

For a given average pulse power. For a given data rate.Baseband transmission To transmit information thru physical channels. 2 PCM waveforms (line codes) are used for binary symbols (M=2). Each waveform carries a symbol from a set of size M. M-ary pulse modulation are used for non-binary symbols (M>2). Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 26 : “shiv rpi” . Eg: M-ary PAM. Each transmit symbol represents k =log M bits of the PCM words. M-ary PAM (M>2) requires less bandwidth than binary PCM. PCM sequences (codewords) are transformed to pulses (waveforms). binary PCM is easier to detect than Mary PAM (M>2).

PAM example: Binary vs 8-ary Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 27 : “shiv rpi” .

Rensselaer Polytechnic Institute ‘10’ -3B Shivkumar Kalyanaraman 28 : “shiv rpi” .Example of M-ary PAM Assuming real time Tx and equal energy per Tx data bit for binary-PAM and 4-ary PAM: • 4-ary: T=2Tb and Binary: T=Tb 2 2 • Energy per symbol in binary-PAM: A = 10B Binary PAM (rectangular pulse) A. 4-ary PAM (rectangular pulse) 3B ‘1’ T T T T -B B ‘11’ ‘01’ T T ‘00’ ‘0’ -A.

Other PCM waveforms: Examples PCM waveforms category: Nonreturn-to-zero (NRZ) Return-to-zero (RZ) +V -V Phase encoded Multilevel binary +V Manchester -V Miller +V -V 1 0 1 1 0 1 0 1 1 0 NRZ-L Unipolar-RZ +V 0 +V Bipolar-RZ 0 -V 0 Rensselaer Polytechnic Institute T 2T 3T +V Dicode NRZ 0 -V 4T 5T 0 T 2T 3T 4T 5T Shivkumar Kalyanaraman 29 : “shiv rpi” .

PCM waveforms: Selection Criteria Criteria for comparing and selecting PCM waveforms: Spectral characteristics (power spectral density and bandwidth efficiency) Bit synchronization capability Error detection capability Interference and noise immunity Implementation cost and complexity Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 30 : “shiv rpi” .

or bit-) rate: Rb = 1 / Tb [bits/sec] Symbol rate : R = 1 / T [symbols/sec] Rensselaer Polytechnic Institute Rb = mR 31 Shivkumar Kalyanaraman : “shiv rpi” . value to l = log 2 L bits (Data bit duration Tb=Ts/l) Mapping every m = log 2 M data bits to a symbol out of M symbols and transmitting a baseband waveform with duration T Quantizing each sampled value to one of the L levels in quantizer. Information (data. Format Bit stream (Data bits) Pulse waveforms (baseband signals) Sample Quantize Encode Pulse modulate Sampling at rate f s = 1 / Ts (sampling time=Ts) Encoding each q.Summary: Baseband Formatting and transmission Digital info. Analog info. Textual source info.

Receiver Structure & Matched Filter Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 32 : “shiv rpi” .

z (T ) & sample r (t ) Rensselaer Polytechnic Institute Major sources of errors: Thermal noise (AWGN) disturbs the signal in an additive fashion (Additive) has flat spectral density for all frequencies of interest (White) is modeled by Gaussian random process (Gaussian Noise) Inter-Symbol Interference (ISI) Due to the filtering effect of transmitter. symbols are “smeared”. K. channel and receiver. Shivkumar Kalyanaraman 33 : “shiv rpi” . M n(t ) Format ˆ mi Detect Demod.Demodulation and detection Format g i (t ) Bandpass si (t ) Pulse modulate modulate channel transmitted symbol hc (t ) mi estimated symbol M-ary modulation i = 1.

Impact of AWGN Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 34 : “shiv rpi” .

5δ (t − 0.75T ) Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 35 : “shiv rpi” .Impact of AWGN & Channel Distortion hc (t ) = δ (t ) − 0.

Receiver job Demodulation and sampling: Waveform recovery and preparing the received signal for detection: Improving the signal power to the noise power (SNR) using matched filter (project to signal space) Reducing ISI using equalizer (remove channel distortion) Sampling the recovered waveform Detection: Estimate the transmitted symbol based on the received sample Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 36 : “shiv rpi” .

Receiver structure Step 1 – waveform to sample transformation Demodulate & Sample Step 2 – decision making Detect r (t ) z (T ) Frequency down-conversion For bandpass signals Receiving filter Equalizing filter Compensation for channel induced ISI Threshold comparison ˆ mi Received waveform Baseband pulse (possibly distorted) Baseband pulse Sample (test statistic) Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 37 : “shiv rpi” .

Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 38 : “shiv rpi” .Baseband vs Bandpass Bandpass model of detection process is equivalent to baseband model because: The received bandpass waveform is first transformed to a baseband waveform. … … yields the same results as … … heterodying the bandpass signal to the baseband . Equivalence theorem: Performing bandpass linear signal processing followed by heterodying the signal to the baseband. followed by a baseband linear signal processing.

we focus on designing a receiver which maximizes the SNR: matched filter Shivkumar Kalyanaraman 39 : “shiv rpi” Rensselaer Polytechnic Institute .Steps in designing the receiver Find optimum solution for receiver design with the following goals: 1. Maximize SNR: matched filter 2. Minimize ISI: equalizer Steps in design: Model the received signal Find separate solutions for each of the goals. First.

Receiver filter to maximize the SNR Model the received signal si (t ) hc (t ) r (t ) n(t ) AWGN r (t ) = si (t ) ∗h c (t ) + n(t ) Simplify the model (ideal channel assumption): Received signal in AWGN Ideal channels hc (t ) = δ (t ) si (t ) n(t ) AWGN r (t ) r (t ) = si (t ) + n(t ) Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 40 : “shiv rpi” .

is the Matched filter.. M is transmitted... given by Matched Filter Receiver h ( t ) = hopt ( t ) = s i (T − t ) * H ( f ) = H opt ( f ) = S i ( f ) exp( − j 2π fT ) * si (t ) which is the time-reversed and delayed version of the conjugate of the transmitted signal h(t ) = hopt (t ) 0 T t 0 T t Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 41 : “shiv rpi” . Solution: The optimum filter.Problem: Design the receiver filter h(t ) such that the SNR is maximized at the sampling time when si (t ).. i = 1.

i.Correlator Receiver The matched filter output at the sampling time. i. s ( t ) > * 0 T Recall: correlation operation is the projection of the received signal onto the signal space! Key idea: Reject the noise (N) outside this space as irrelevant: => maximize S/N Shivkumar Kalyanaraman 42 : “shiv rpi” Rensselaer Polytechnic Institute . correlation or inner product! z (T ) = hopt (T ) ∗ r (T ) = ∫ r (τ )s i (τ ) d τ =< r ( t ). can be realized as the correlator output. Matched filtering.e.e. convolution with si*(T-τ) simplifies to integration w/ si*(τ).

Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 43 : “shiv rpi” .Irrelevance Theorem: Noise Outside Signal Space Noise PSD is flat (“white”) => total noise power infinite across the spectrum. We care only about the noise projected in the finite signal dimensions (eg: the bandwidth of interest).

This is like constructive interference.where it has significant value . τ ) between them When the two signals are similar in shape and unshifted with respect to each other. Correlation is a measure of the similarity between two signals as a function of time shift (“lag”.Aside: Correlation Effect Correlation is a maximum when two signals are similar in shape. Rensselaer Polytechnic Institute Shivkumar Kalyanaraman : “shiv rpi” 44 .shows for how long the signals remain similar. The breadth of the correlation function . and are in phase (or 'unshifted' with respect to each other). their product is all positive.

Aside: Autocorrelation Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 45 : “shiv rpi” .

Aside: Cross-Correlation &Radar Figure: shows how the signal can be located within the noise. A copy of the known reference signal is correlated with the unknown signal. Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 46 : “shiv rpi” . The large value of the correlation indicates when the reference signal occurs. A large value for correlation shows the degree of confidence that the reference signal is detected. The correlation will be high when the reference is similar to the unknown signal.

and added noise Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 47 : “shiv rpi” Source: Bores Signal Processing . • The unknown signal is correlated with a number of known reference functions. • The largest value for correlation is the most likely match. • The correlation will be high if the reference is similar to the unknown signal. • Same principle in communications: reference signals corresponding to symbols • The ideal communications channel may have attenuated.• A copy of a known reference signal is correlated with the unknown signal. phase shifted the reference signal. • A large value for correlation shows the degree of similarity to the reference.

and the transmitted signal vector as s Matched filter “projects” the r onto signal space spanned by s (“matches” it) Filtered signal can now be safely sampled by the receiver at the correct sampling instants.Matched Filter: back to cartoon… Consider the received signal as a vector r. resulting in a correct interpretation of the binary message Matched filter is the filter that maximizes the signal-to-noise ratio it can be shown that it also minimizes the BER: it is a simple projection operation Shivkumar Kalyanaraman Rensselaer Polytechnic Institute 48 : “shiv rpi” .

Example of matched filter (real signals) si (t ) A T h opt (t ) A T y (t ) = si (t ) ∗h opt (t ) A2 T t T t 0 T 2T t si (t ) A T h opt (t ) A T y (t ) = si (t ) ∗h opt (t ) A2 T/2 T −A T t −A T T/2 T t 0 T/2 T 3T/2 2T t − A2 2 Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 49 : “shiv rpi” .

Z ( f ) =| S ( f ) |2 exp(− j 2πfT ) 2. The Fourier transform of a matched filter output with the matched signal as input is. spectral amplitude matching that gives optimum SNR to the peak value. proportional to the ESD of the input signal. z (t ) = Rs (t − T ) ⇒ z (T ) = Rs (0) = Es 3. The output signal of a matched filter is proportional to a shifted version of the autocorrelation function of the input signal to which the filter is matched. Shivkumar Kalyanaraman 50 : “shiv rpi” Rensselaer Polytechnic Institute . Two matching conditions in the matched-filtering operation: spectral phase matching that gives the desired output peak at time T. The output SNR of a matched filter depends only on the ratio of the signal energy to the PSD of the white noise at the filter input. except for a time delay factor.Properties of the Matched Filter 1. Es ⎛S⎞ max⎜ ⎟ = ⎝ N ⎠T N 0 / 2 4.

.. z M (T )) = ( z1 .... M z = ( z1 (T ).... z 2 (T ). z 2 ... z M ) Note: we are projecting along the basis directions of the signal space Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 51 : “shiv rpi” .Implementation of matched filter receiver Bank of M matched filters s (T − t ) * 1 z1 (T ) r (t ) sM (T − t ) * zM ⎡ z1 ⎤ ⎢ M ⎥=z ⎢ ⎥ ⎢zM ⎥ ⎣ ⎦ (T ) Matched filter output: z Observation vector zi = r (t ) ∗ s ∗i (T − t ) i = 1...

M Note: In previous slide we “filter” i.. z M ) T i = 1... 52 : “shiv rpi” .Implementation of correlator receiver Bank of M correlators s ∗1 (t ) r (t ) s ∗ M ∫ (t ) T z1 (T ) 0 ∫ zi = ∫ r (t )si (t )dt 0 Rensselaer Polytechnic Institute T 0 z M (T ) ⎡ z1 ⎤ ⎢M⎥ ⎢ ⎥ ⎢zM ⎥ ⎣ ⎦ =z Correlators output: z Observation vector z = ( z1 (T )...e. z 2 .... convolute in the boxesKalyanaraman Shivkumar shown.... z M (T )) = ( z1 . z 2 (T )..

Implementation example of matched filter receivers s1 (t ) A T Bank of 2 matched filters 0 T t A T z1 (T ) T r (t ) s2 (t ) 0 T t 0 0 −A T −A T T z 2 (T ) ⎡z1 ⎤ ⎢ ⎥ ⎢ ⎥ ⎢z2 ⎥ ⎣ ⎦ =z z Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 53 : “shiv rpi” .

but misses some signal power Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 54 : “shiv rpi” .Matched Filter: Frequency domain View Simple Bandpass Filter: excludes noise.

Matched Filter: Frequency Domain View (Contd) Multi-Bandpass Filter: includes more signal power. but adds more noise also! Matched Filter: includes more signal power. weighted according to size => maximal noise rejection! Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 55 : “shiv rpi” .

Maximal Ratio Combining (MRC) viewpoint Generalization of this f-domain picture. for combining multi-tap signal Weight each branch SNR: Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 56 : “shiv rpi” .

Examples of matched filter output for bandpass modulation schemes Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 57 : “shiv rpi” .

Signal Space Concepts Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 58 : “shiv rpi” .

Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 59 : “shiv rpi” .Signal space: Overview What is a signal space? Vector representations of signals in an N-dimensional orthogonal space Why do we need a signal space? It is a means to convert signals to vectors and vice versa. It is a means to calculate signals energy and Euclidean distances between signals. Why are we interested in Euclidean distances between signals? For detection purposes: The received signal is transformed to a received vectors. The signal which has the minimum distance to the received signal is estimated as the transmitted signal.

a32 ) z (t ) = z1ψ 1 (t ) + z2ψ 2 (t ) ⇔ z = ( z1 . z 2 ) 60 Shivkumar Kalyanaraman : “shiv rpi” .Schematic example of a signal space ψ 2 (t ) s1 = (a11 . a12 ) ψ 1 (t ) z = ( z1 . a22 ) Transmitted signal alternatives Received signal at matched filter output Rensselaer Polytechnic Institute s1 (t ) = a11ψ 1 (t ) + a12ψ 2 (t ) ⇔ s1 = (a11 . a22 ) s3 (t ) = a31ψ 1 (t ) + a32ψ 2 (t ) ⇔ s 3 = (a31 . a12 ) s2 (t ) = a21ψ 1 (t ) + a22ψ 2 (t ) ⇔ s 2 = (a21 . a32 ) s 2 = (a21 . z 2 ) s 3 = (a31 .

y (t ) >= −∞ x(t ) y * (t )dt ∫ = cross-correlation between x(t) and y(t) Properties of inner product: < ax (t ). y (t ) > < x(t ) + y (t ). ay (t ) >= a * < x(t ). y (t ) >= a < x(t ).Signal space To form a signal space. y (t ) > < x(t ). z (t ) > Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 61 : “shiv rpi” . z (t ) >=< x(t ). z (t ) > + < y (t ). first we need to know the inner product between two signals (functions): Inner (scalar) product: ∞ < x(t ).

What is norm? Norm of a signal: x(t ) = < x(t ). y = x(t ) − y (t ) We refer to the norm between two signals as the Euclidean distance between two signals. Rensselaer Polytechnic Institute Shivkumar Kalyanaraman : “shiv rpi” 62 . x(t ) > = = “length” of x(t) ∫ ∞ −∞ x(t ) dt = E x 2 ax(t ) = a x(t ) Norm between two signals: d x .Signal space … The distance in signal space is measure by calculating the norm.

**Example of distances in signal space
**

ψ 2 (t )

s1 = (a11 , a12 )

E1

E3

d s1 , z

ψ 1 (t )

s 3 = (a31 , a32 )

d s3 , z

z = ( z1 , z 2 )

E2

d s2 , z

s 2 = (a21 , a22 )

The Euclidean distance between signals z(t) and s(t):

d si , z = si (t ) − z (t ) = (ai1 − z1 ) 2 + (ai 2 − z 2 ) 2 i = 1,2,3

Rensselaer Polytechnic Institute

Shivkumar Kalyanaraman 63

: “shiv rpi”

**Orthogonal signal space
**

N-dimensional orthogonal signal space is characterized by N N ψ linearly independent functions { j (t )}j =1 called basis functions. The basis functions must satisfy the orthogonality condition

< ψ i (t ),ψ j (t ) >= ∫ψ i (t ) * (t )dt = K iδ ji ψj

0

T

0≤t ≤T j , i = 1,..., N

where

δ ij = ⎨

⎧1 → i = j ⎩0 → i ≠ j

If all K i = 1 , the signal space is orthonormal. Constructing Orthonormal basis from non-orthonormal set of vectors: Gram-Schmidt procedure

Rensselaer Polytechnic Institute

Shivkumar Kalyanaraman 64

: “shiv rpi”

**Example of an orthonormal bases
**

Example: 2-dimensional orthonormal signal space

⎧ 2 cos(2πt / T ) ⎪ψ 1 (t ) = T ⎪ ⎨ ⎪ψ (t ) = 2 sin( 2πt / T ) ⎪ 2 T ⎩

T

0≤t <T 0≤t <T

ψ 2 (t )

< ψ 1 (t ),ψ 2 (t ) >= ∫ψ 1 (t ) 2 (t )dt = 0 ψ

0

0

ψ 1 (t )

ψ 1 (t ) = ψ 2 (t ) = 1

**Example: 1-dimensional orthonornal signal space
**

ψ 1 (t )

1 T

ψ 1 (t ) = 1

0 T t 65

ψ 1 (t )

Shivkumar Kalyanaraman

: “shiv rpi”

0

Rensselaer Polytechnic Institute

**Sine/Cosine Bases: Note!
**

Approximately orthonormal!

**These are the in-phase & quadrature-phase dimensions of complex baseband equivalent representations.
**

Rensselaer Polytechnic Institute

Shivkumar Kalyanaraman 66

: “shiv rpi”

but only one dimension in BPSK.Example: BPSK Note: two symbols. Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 67 : “shiv rpi” .

......... M N≤M T where 1 1 * aij = < si (t ). N 0≤t ≤T i = 1....(t )}N ψ N≤M j j =1 Signal space … si (t ) = ∑ aijψ j (t ) j =1 N i = 1.Any arbitrary finite set of waveforms {si (t )}iM1 = where each member of the set is of duration T. can be expressed as a linear combination of N orthogonal waveforms where { ..ψ j (t ) >= ∫ si (t )ψ j (t )dt Kj Kj 0 j = 1... M s i = (ai1 . aiN ) Vector representation of waveform Rensselaer Polytechnic Institute Ei = ∑ K j aij j =1 N 2 Waveform energy Shivkumar Kalyanaraman 68 : “shiv rpi” . ai 2 ..

aiN ) Vector to waveform conversion Waveform to vector conversion ψ 1 (t ) si (t ) ∫ ψ N (t ) T ai1 ψ 1 (t ) 0 ∫ Rensselaer Polytechnic Institute T 0 aiN ⎡ai1 ⎤ ⎢ M ⎥ = sm ⎢ ⎥ ⎢aiN ⎥ ⎣ ⎦ sm ⎡ai1 ⎤ ⎢M⎥ ⎢ ⎥ ⎢aiN ⎥ ⎣ ⎦ ai1 ψ N (t ) si (t ) aiN Shivkumar Kalyanaraman 69 : “shiv rpi” ... ai 2 .Signal space … s i (t ) = ∑ a ijψ j (t ) j =1 N s i = (ai1 ...

.. a12 ) s2 (t ) = a21ψ 1 (t ) + a22ψ 2 (t ) ⇔ s 2 = (a21 .. a22 ) s3 (t ) = a31ψ 1 (t ) + a32ψ 2 (t ) ⇔ s 3 = (a31 . M Shivkumar T 0 ≤ t ≤ Kalyanaraman : “shiv rpi” .. N 70 i = 1.. a22 ) Transmitted signal alternatives s1 (t ) = a11ψ 1 (t ) + a12ψ 2 (t ) ⇔ s1 = (a11 . a32 ) s 2 = (a21 . a32 ) aij = ∫ si (t )ψ j (t )dt 0 T Rensselaer Polytechnic Institute j = 1.Example of projecting signals to an orthonormal signal space ψ 2 (t ) s1 = (a11 .. a12 ) ψ 1 (t ) s 3 = (a31 ...

z 2 ..Matched filter receiver (revisited) (note: we match to the basis directions) Bank of N matched filters ψ (T − t ) ∗ 1 z1 r (t ) ψ ∗ N (T − t ) zN ⎡ z1 ⎤ ⎢ ⎥=z ⎢ ⎥ ⎢zN ⎥ ⎣ ⎦ Observation vector z si (t ) = ∑ aijψ j (t ) N z = ( z1 . z N ) z j = r (t ) ∗ψ j (T − t ) Rensselaer Polytechnic Institute j =1 i = 1..... N Shivkumar Kalyanaraman 71 : “shiv rpi” .... M N≤M j = 1.....

.. N Shivkumar Kalyanaraman 72 : “shiv rpi” .. z 2 ... z N ) 0 Rensselaer Polytechnic Institute j =1 z j = ∫ r (t ) j (t )dt ψ T N≤M j = 1.... M z = ( z1 ....Correlator receiver (revisited) Bank of N correlators ψ 1 (t ) r (t ) ψ N (t ) ∫ ∫ T z1 0 T 0 zN ⎡ r1 ⎤ ⎢M⎥ ⎢ ⎥ ⎢rN ⎥ ⎣ ⎦ =z z Observation vector si (t ) = ∑ aijψ j (t ) N i = 1..

Example of matched filter receivers using basic functions s1 (t ) A T s2 (t ) 0 0 T t −A T ψ 1 (t ) 1 T T t 0 T t 1 matched filter ψ 1 (t ) r (t ) 1 T z1 T t [z1] =z z 0 Number of matched filters (or correlators) is reduced by 1 compared to using matched filters (correlators) to the transmitted signal! Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 73 : “shiv rpi” .

**White noise in Orthonormal Signal Space
**

AWGN n(t) can be expressed as

~ ˆ n(t ) = n(t ) + n (t )

Noise projected on the signal space (colored):impacts the detection process. Noise outside on the signal space (irrelevant)

ˆ n(t ) = ∑ n jψ j (t )

N

n j =< n(t ),ψ j (t ) > ~ < n (t ),ψ (t ) >= 0

j

j =1

Vector representation of

ˆ n(t )

j = 1,..., N j = 1,..., N

**independent zero-mean Gaussain random variables with variance var(n j ) = N 0 / 2
**

j j =1

{n }

n = (n1 , n2 ,..., nN )

N

Rensselaer Polytechnic Institute

Shivkumar Kalyanaraman 74

: “shiv rpi”

Detection: Maximum Likelihood & Performance Bounds

Rensselaer Polytechnic Institute

Shivkumar Kalyanaraman 75

: “shiv rpi”

**Detection of signal in AWGN
**

Detection problem: Given the observation vector z , perform a mapping from z ˆ to an estimate m of the transmitted symbol, mi , such that the average probability of error in the decision is minimized.

n

mi

Modulator

si

z

Decision rule

ˆ m

Rensselaer Polytechnic Institute

Shivkumar Kalyanaraman 76

: “shiv rpi”

...Statistics of the observation Vector AWGN channel model: z = si + n Signal vector s i = (ai1 .. Its pdf is ⎛ z − si 2 ⎞ 1 ⎟ pz ( z | s i ) = exp⎜ − (πN 0 )N / 2 ⎜ N 0 ⎟ ⎝ ⎠ Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 77 : “shiv rpi” .... Elements of noise vector n = (n1 . n2 .. The noise vector pdf is ⎛ n2⎞ 1 ⎟ pn (n) = exp⎜ − (πN 0 )N / 2 ⎜ N 0 ⎟ ⎝ ⎠ The elements of observed vector z = ( z1 .i. ai 2 .. z N ) are independent Gaussian random variables.. nN ) are i.d Gaussian random variables with zero-mean and variance N 0 / 2 . z 2 ... aiN ) is deterministic.

.. is maximum for all k = i pz ( z ) Shivkumar Kalyanaraman 78 : “shiv rpi” Rensselaer Polytechnic Institute ...Detection Optimum decision rule (maximum a posteriori probability): ˆ Set m = mi if Pr( mi sent | z ) ≥ Pr(mk sent | z ). Applying Bayes’ rule gives: ˆ Set m = mi if pk pz (z | mk ) . for all k ≠ i where k = 1. M .

..Detection … Partition the signal space into M decision regions.. Z M Vector z lies inside region Z i if ln[ pk pz (z | mk ) ]. such that Z1 . is maximum for all k = i.. pz ( z ) That means ˆ m = mi Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 79 : “shiv rpi” .

Detection (ML rule) For equal probable symbols. is maximum for all k = i or equivalently: ˆ Set m = mi if ln[ pz (z | mk )]. Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 80 : “shiv rpi” . is maximum for all k = i which is known as maximum likelihood. the optimum decision rule (maximum posteriori probability) is simplified to: ˆ Set m = mi if pz (z | mk ).

is minimum for all k = i Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 81 : “shiv rpi” ..Detection (ML)… Partition the signal space into M decision regions.. is maximum for all k = i That means ˆ m = mi Vector z lies inside region Z i if z − s k .. Z1 .. Z M Restate the maximum likelihood decision rule as follows: Vector z lies inside region Z i if ln[ pz (z | mk )].

Schematic example of ML decision regions ψ 2 (t ) Z2 s2 s3 Z3 s1 Z1 ψ 1 (t ) s4 Z4 Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 82 : “shiv rpi” .

an error in decision occurs if the observation vector z does not fall inside region Z i.Probability of symbol error Erroneous decision: For the transmitted symbol s i or equivalently signal vector mi . Probability of erroneous decision for a transmitted symbol ˆ Pr(m ≠ mi ) = Pr(mi sent)Pr(z does not lie inside Z i mi sent) Probability of correct decision for a transmitted symbol ˆ Pr(m = mi ) = Pr(mi sent)Pr(z lies inside Z i mi sent) Pc (mi ) = Pr(z lies inside Z i mi sent) = Pe (mi ) = 1 − Pc (mi ) Rensselaer Polytechnic Institute Zi ∫ p (z | m )dz z i Shivkumar Kalyanaraman 83 : “shiv rpi” .

Example for binary PAM pz (z | m2 ) pz (z | m1 ) s2 − Eb 0 s1 Eb ψ 1 (t ) ⎛ s1 − s 2 / 2 ⎞ ⎟ Pe (m1 ) = Pe (m2 ) = Q⎜ ⎜ N /2 ⎟ 0 ⎝ ⎠ ⎛ 2 Eb PB = PE (2) = Q⎜ ⎜ N 0 ⎝ Rensselaer Polytechnic Institute ⎞ ⎟ ⎟ ⎠ Shivkumar Kalyanaraman : “shiv rpi” 84 .

of symbol error … Average probability of symbol error : PE ( M ) = ˆ ∑ Pr ( m ≠ m ) i =1 i M For equally probable symbols: 1 PE ( M ) = M 1 ∑ Pe (mi ) = 1 − M i =1 M i =1 Z i z i M ∑ P (m ) i =1 c i M 1 = 1− M Rensselaer Polytechnic Institute ∑ ∫ p (z | m )dz Shivkumar Kalyanaraman 85 : “shiv rpi” .Average prob.

si ) Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 86 : “shiv rpi” .Union bound Union bound The probability of a finite union of events is upper bounded by the sum of the probabilities of the individual events. s i ) k =1 k ≠i M 1 PE ( M ) ≤ M ∑∑ P (s i =1 k =1 k ≠i 2 M M k . Pe (mi ) ≤ ∑ P2 (s k .

s1 ) = ∫ pr (r | m1 )dr A4 Shivkumar Kalyanaraman : “shiv rpi” 87 . s1 ) = 2 Rensselaer Polytechnic Institute A ∫ p (r | m )dr r 1 P2 (s 3 . s1 ) = ∫ p (r | m )dr r 1 P2 (s 4 .Example of union bound Pe (m1 ) = r Z 2 ∪Z3 ∪Z 4 ∫ p (r | m )dr 1 Z2 s2 r ψ2 Z1 s1 Union bound: Pe (m1 ) ≤ ∑ P2 (s k . s1 ) k =2 4 ψ1 Z3 s3 s4 Z4 A2 r s2 ψ2 s1 s2 r ψ1 ψ2 s1 s2 r ψ1 ψ2 s1 ψ1 s3 s3 s4 s3 A3 A3 s4 A4 s4 P2 (s 2 .

Upper bound based on minimum distance P2 (s k . si ) ≤ (M − 1)Q⎜ N / 2 ⎟ ⎜ ⎟ i =1 k =1 0 ⎝ ⎠ k ≠i M M Minimum distance in the signal space: d min = min d ik i . s i ) = Pr(z is closer to s k than s i .k i≠k Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 88 : “shiv rpi” . when s i is sent) ∞ = d ik ∫ ⎛ d ik / 2 u2 1 exp(− )du =Q⎜ ⎜ N /2 N0 πN 0 0 ⎝ d ik = s i − s k ⎞ ⎟ ⎟ ⎠ 1 PE ( M ) ≤ M ⎛ d min / 2 ⎞ ∑∑ P2 (s k .

4 Es ψ 1 (t ) s4 − Es Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 89 : “shiv rpi” . 4 s1 d1. based on union bound s i = Ei = Es ..k = 2 Es i≠k d min = 2 Es s3 − Es Es s2 d1. Symbol error prob... 2 d 2. i = 1.Example of upper bound on av..3 d 3.4 ψ 2 (t ) d i .

Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 90 : “shiv rpi” .

are energy signal (zero power). Eb STb S W = = N 0 N / W N Rb Rensselaer Polytechnic Institute Rb W : Bit rate : Bandwidth Note: S/N = Eb/No x spectral efficiency 91 Shivkumar Kalyanaraman : “shiv rpi” . because: Signals are transmitted within a symbol duration and hence. SNR should be modified in terms of bit-energy in digital communication. A metric at the bit-level facilitates comparison of different DCS transmitting different number of bits per symbol.Eb/No figure of merit in digital communications SNR or S/N is the average signal power to the average noise power.

Example of Symbol error prob. For PAM signals Binary PAM s2 s1 − Eb 0 ψ 1 (t ) Eb s4 −6 Eb 5 −2 4-ary PAM s3 s2 Eb 5 s1 6 Eb 5 ψ 1 (t ) 0 2 Eb 5 ψ1 (t ) 1 T 0 Rensselaer Polytechnic Institute T t : “shiv rpi” Shivkumar Kalyanaraman 92 .

uB is a “sufficient statistic”. Noise outside these finite dimensions is irrelevant for detection.Maximum Likelihood (ML) Detection: Vector Case Nearest Neighbor Rule: By the isotropic property of the Gaussian noise. Error probability: Project the received vector y along the difference vector direction uA. we expect the error probability to be the same for both the transmit symbols uA. uB. (rotational invariance of detection problem) ps: Vector norm is a natural extension of “magnitude” or length Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 93 : “shiv rpi” .

Extension to M-PAM (Multi-Level Modulation) Note: h refers to the constellation shape/direction MPAM: Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 94 : “shiv rpi” .

use v* for complex vectors (“transpose and conjugate”) for inner products… Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 95 : “shiv rpi” .Complex Vector Space Detection Error probability: Note: Instead of vT.

Complex Detection: Summary Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 96 : “shiv rpi” .

Detection Error => BER If the bit error is i.d (discrete memoryless channel) over the sequence of bits. then you can model it as a binary symmetric channel (BSC) BER is modeled as a uniform probability f As BER (f) increases.i. the effects become increasingly intolerable f tends to increase rapidly with lower SNR: “waterfall” curve (Q-function) Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 97 : “shiv rpi” .

higher bandwidth (upto 10 Mhz. high bandwidth ! But cant process w/ complicated codes. and 20Mhz LAN) Optical fiber comm: High SNR.7kHz) Wireless: Low SNR = 5-10dB. signal processing etc Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 98 : “shiv rpi” .SNR vs BER: AWGN vs Rayleigh Need diversity techniques to deal with Rayleigh (even 1-tap. MAN. flat-fading)! Observe the “waterfall” like characteristic (essentially plotting the Q(x) function)! Telephone lines: SNR = 37dB. but low b/w (3.

The multiple signal copies should experience uncorrelated fading in the channel. As a rough rule: Pe is proportional to BER BER 1 γ 0L Diversity of Diversity of L:th order L:th order Average SNR Average SNR Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 99 : “shiv rpi” . In this case the probability that all signal copies fade simultaneously is reduced dramatically with respect to the probability that a single copy experiences a fade.Better performance through diversity Diversity the receiver is provided with multiple copies of the transmitted signal.

Rayleigh fading. L=1 AWGN channel (no fading) SNR L=4 L=3 L=2 (= γ0) We will explore this story later… slide set part II Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 100 : “shiv rpi” . SNR (diversity effect) BER ( = Pe ) Flat fading channel.BER vs.

Modulation Techniques Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 101 : “shiv rpi” .

What is Modulation? Encoding information in a manner suitable for transmission. phase or frequency of a carrier Demodulation: extract baseband message from carrier Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 102 : “shiv rpi” . Translate baseband source signal to bandpass signal Bandpass signal: “modulated signal” How? Vary amplitude.

modulation. better security: CDMA. impairment resistance: Using coding. diversity Equalization. power error correction. more power efficient Higher data rates. multicarrier techniques for ISI mitigation More efficient multiple access strategies.Digital vs Analog Modulation Cheaper. encryption etc Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 103 : “shiv rpi” . faster.

Goals of Modulation Techniques • High Bit Rate • High Spectral Efficiency (max Bps/Hz) • High Power Efficiency (min power to achieve a target BER) • Low-Cost/Low-Power Implementation • Robustness to Impairments Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 104 : “shiv rpi” .

Modulation: representation • Any modulated signal can be represented as s(t) = A(t) cos [ωct + φ(t)] amplitude phase or frequency . target BER at lower SNRs) Shivkumar Kalyanaraman 105 : “shiv rpi” .A(t) sin φ(t) sin ωct quadrature s(t) = A(t) cos φ(t) cos ωct in-phase • Linear versus nonlinear modulation ⇒ impact on spectral efficiency Linear: Amplitude or phase Non-linear: frequency: spectral broadening • Constant envelope versus non-constant envelope ⇒ hardware implications with impact on power efficiency Rensselaer Polytechnic Institute (=> reliability: i.e.

Complex Vector Spaces: Constellations MPSK Circular Square Each signal is encoded (modulated) as a vector in a signal space Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 106 : “shiv rpi” .

π/4-DQPSK) Shivkumar Kalyanaraman 107 : “shiv rpi” . quadrature LINEAR MODULATIONS Square Constellations M-ARY QUADRATURE AMPLITUDE MOD. (M-QAM) M-ARY PHASE SHIFT KEYING (M-PSK) Circular Constellations M≠4 M=4 (4-QAM = 4-PSK) M≠4 CONVENTIONAL 4-PSK (QPSK) Rensselaer Polytechnic Institute OFFSET DIFFERENTIAL 4-PSK 4-PSK (OQPSK) (DQPSK. in-phase Q(t).Linear Modulation Techniques s(t) = [ Σ an g (t-nT)]cos ωct .[ Σ bn g (t-nT)] sin ωct n n I(t).

– M-QAM is more spectrally efficient than M-PSK but also more sensitive to system nonlinearities. Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 108 : “shiv rpi” .M-PSK and M-QAM M-PSK (Circular Constellations) bn 4-PSK 16-PSK an M-QAM (Square Constellations) bn 16-QAM 4-PSK an Tradeoffs – Higher-order modulations (M large) are more spectrally efficient but less power efficient (i. BER higher).e.

Bandwidth vs. Power Efficiency MPSK: MQAM: MFSK: Rensselaer Polytechnic Institute Source: Rappaport book. chap 6 Shivkumar Kalyanaraman 109 : “shiv rpi” .

Make sure that the neighboring symbol has only 1-bit difference (hamming distance = 1) Rensselaer Polytechnic Institute Binary PAM s2 s1 − Eb 0 ψ 1 (t ) Eb s4 −6 Eb 5 −2 4-ary PAM s3 s2 Eb 5 s1 6 Eb 5 ψ 1 (t ) 0 2 Eb 5 Gray coding 00 s4 −6 Eb 5 −2 01 11 10 s1 6 Eb 5 ψ 1 (t ) 4-ary PAM s3 s2 Eb 5 0 2 Eb 5 Shivkumar Kalyanaraman 110 : “shiv rpi” .MPAM & Symbol Mapping Note: the average energy per-bit is constant Gray coding used for mapping bits to symbols Why? Most likely error is to confuse with neighboring symbol.

MPAM: Details Unequal energies/symbol: Rensselaer Polytechnic Institute Shivkumar Kalyanaraman Decision111 Regions : “shiv rpi” .

M-PSK (Circular Constellations) MPSK: bn 4-PSK 16-PSK an Constellation points: Equal energy in all signals: 01 01 00 00 Gray coding 11 Rensselaer Polytechnic Institute 10 10 11 Shivkumar Kalyanaraman 112 : “shiv rpi” .

Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 113 : “shiv rpi” .MPSK: Decision Regions & Demod’ln Z2 Z3 Z1 Z5 Z6 Z8 Z4 Z3 Z2 Z1 Z4 Z7 4PSK 4PSK: 1 bit/complex dimension or 2 bits/symbol 8PSK m=1 si(t) + n(t) X g(Tb .t) Z1: r > 0 Z2: r ≤ 0 m=0 m = 0 or 1 cos(2πfct) Coherent Demodulator for BPSK.

M-QAM (Square Constellations) MQAM: bn 16-QAM 4-PSK an Unequal symbol energies: MQAM with square constellations of size L2 is equivalent to MPAM modulation with constellations of size L on each of the in-phase and quadrature signal components For square constellations it takes approximately 6 dB more power to send an additional 1 bit/dimension or 2 bits/symbol while maintaining the same minimum distance between constellation points Hard to find a Gray code mapping where all adjacent symbols differ by a single bit Rensselaer Polytechnic Institute Z1 Z5 Z9 Z13 Z2 Z6 Z10 Z14 Z3 Z7 Z11 Z15 Z4 Z8 Z12 Z16 16QAM: Decision Regions Shivkumar Kalyanaraman : “shiv rpi” 114 .

phase of the transmitted signal carrier φ0 must be matched to the phase of the receiver carrier φ More cost. more sensitive to doppler effects: decorrelation of signal phase in time-domain Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 115 : “shiv rpi” .Non-Coherent Modulation: DPSK Information in MPSK. MQAM carried in signal phase. susceptible to carrier phase drift. Harder to obtain in fading channels Differential modulation: do not require phase reference.e. More general: modulation w/ memory: depends upon prior symbols transmitted. Use prev symbol as the a phase reference for current symbol Info bits encoded as the differential phase between current & previous symbol Less sensitive to carrier phase drift (f-domain) . Requires coherent demodulation: i.

If symbol over time [(k−1)Ts. whereas a 1 bit is encoded as a phase change of π. then to encode a 0 bit over [kTs. π. kTs) has phase θ(k − 1) = ejθi .Differential Modulation (Contd) DPSK: Differential BPSK A 0 bit is encoded by no change in phase. DQPSK: gray coding: Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 116 : “shiv rpi” . the symbol would have phase: θ(k) = ejθi and… … to encode a 1 bit the symbol would have phase θ(k) = ej(θi+π). (k + 1)Ts). θi = 0.

. where the O indicates the offset QPSK modulation with quadrature offset is referred to as O-QPSK O-QPSK has the same spectral properties as QPSK for linear amplification. Abrupt phase transitions and large amplitude variations can be distorted by nonlinear amplifiers and filters Avoided by offsetting the quadrature branch pulse g(t) by half a symbol period Usually abbreviated as O-MPSK. … but has higher spectral efficiency under nonlinear amplification. since the maximum phase transition of the signal is 90 degrees Another technique to mitigate the amplitude fluctuations of a 180 degree phase shift used in the IS-54 standard for digital cellular is π/4-QPSK Maximum phase transition of 135 degrees.. versus 90 degrees for offset QPSK and 180 degrees for QPSK Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 117 : “shiv rpi” .Quadrature Offset Phase transitions of 180o can cause large amplitude transitions (through zero point).

Offset QPSK waveforms Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 118 : “shiv rpi” .

Frequency Shift Keying (FSK) • Continuous Phase FSK (CPFSK) – digital data encoded in the frequency shift – typically implemented with frequency modulator to maintain continuous phase s(t) = A cos [ωct + 2 πkf ∫ d(τ) dτ] t −∞ – nonlinear modulation but constant-envelope • Minimum Shift Keying (MSK) – minimum bandwidth. and HIPERLAN Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 119 : “shiv rpi” . CDPD. sidelobes large – can be implemented using I-Q receiver • Gaussian Minimum Shift Keying (GMSK) – reduces sidelobes of MSK using a premodulation filter – used by RAM Mobile Data.

Minimum Shift Keying (MSK) spectra Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 120 : “shiv rpi” .

16 0.0 2.5 2.5 1.0 -100 -120 0 (MSK) 0.25 -80 1.5 Shivkumar Kalyanaraman : “shiv rpi” Normalized Frequency (f-fc)Tb Rensselaer Polytechnic Institute 121 .0 1.Spectral Characteristics Power Spectral Density (dB) 10 0 -20 -40 -60 QPSK/DQPSK GMSK B3-dBTb = 0.

QPSK Pb 2 10-4 5 2 10-5 5 2 10-6 0 2 4 6 8 10 12 14 DQPSK BPSK 6.5 dB QPSK 6. is more spectrally efficient but requires more SNR per bit. SNR/bit.5 dB DBPSK ~8 dB DQPSK ~9 dB • QPSK is more spectrally efficient than BPSK with the same performance. • There is ~3 dB power penalty for differential detection. for M>4. • M-PSK.10-1 5 2 10-2 5 2 10-3 5 Bit Error Probability (BER): AWGN -3 For Pb = 10 DBPSK BPSK. γb. dB Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 122 : “shiv rpi” .

Bit Error Probability (BER): Fading Channel 1 5 2 10-1 5 2 10-2 5 DBPSK Pb • Pb is inversely proportion to the average SNR per bit. 2 10-3 5 2 10-4 5 2 10-5 0 5 10 15 20 25 30 35 AWGN BPSK • Transmission in a fading environment requires about 18 dB more power for Pb = 10-3. SNR/bit. dB Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 123 : “shiv rpi” . γb.

dB Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 124 : “shiv rpi” .Bit Error Probability (BER): Doppler Effects • Doppler causes an irreducible error floor when differential detection is used ⇒ decorrelation of reference signal. data rate T 10-4s 10-5s 10-6s Pbfloor 3x10-4 3x10-6 3x10-8 10 kbps 100 kbps 1 Mbps -3 Pb 10 fDT=0. SNR/bit. For fD = 80 Hz.001 10 -5 0 10 -6 0 10 20 30 40 50 60 The implication is that Doppler is not an issue for high-speed wireless data. CRC Press.002 0. Foundations of Mobile Radio Engineering . Yacoub. [M. 1993] γb. 100 QPSK DQPSK 10 -1 Rayleigh Fading 10 -2 • The irreducible Pb depends on the data rate and the Doppler. D.003 10 -4 No Fading 0.

τ Maximum Bit Rate Mobile (rural) 25 μsec 8 kbps Mobile (city) 2." IEEE JSAC. "The Effects of Time Delay Spread on Portable Radio Communications Channels with Digital Modulation.-I. for QPSK. For example. 10-1 Coherent Detection + BPSK QPSK OQPSK Modulation x MSK Irreducible Pb 10-2 x x x x + + + + x + • The rms delay spread imposes a limit on the maximum bit rate in a multipath environment. June 1987] 10-3 10-4 10-2 10-1 rms delay spread τ = symbol period T 100 Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 125 : “shiv rpi” .Bit Error Probability (BER): Delay Spread • ISI causes an irreducible error floor.5 μsec 80 kbps Microcells 500 nsec 400 kbps Large Building 100 nsec 2 Mbps [J. C. Chuang.

Summary of Modulation Issues • Tradeoffs – linear versus nonlinear modulation – constant envelope versus non-constant envelope – coherent versus differential detection – power efficiency versus spectral efficiency • Limitations – flat fading – doppler – delay spread Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 126 : “shiv rpi” .

Pulse Shaping Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 127 : “shiv rpi” .

Recall: Impact of AWGN only Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 128 : “shiv rpi” .

Impact of AWGN & Channel Distortion hc (t ) = δ (t ) − 0.5δ (t − 0.75T ) Multi-tap. ISI channel Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 129 : “shiv rpi” .

wireless channels) Channels behave like band-limited filters H c ( f ) = H c ( f ) e jθ c ( f ) Non-constant amplitude Amplitude distortion Non-linear phase Phase distortion Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 130 : “shiv rpi” .g.ISI Effects: Band-limited Filtering of Channel ISI due to filtering effect of the communications channel (e.

Inter-Symbol Interference (ISI) ISI in the detection process due to the filtering effects of the system Overall equivalent system transfer function H ( f ) = Ht ( f )H c ( f )H r ( f ) creates echoes and hence time dispersion causes ISI at sampling time ISI effect z k = sk + nk + ∑ α i si i≠k Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 131 : “shiv rpi” .

Inter-symbol interference (ISI): Model {xk } x1 x2 Baseband system model Tx filter ht (t ) Channel hc (t ) r (t ) Rx. filter hr (t ) Hr ( f ) zk t = kT Detector ˆ {xk } T x3 Ht ( f ) T Hc ( f ) n(t ) Equivalent model {xk } x1 x2 Equivalent system h(t ) H( f ) z (t ) zk t = kT Detector ˆ {xk } T x3 T ˆ n(t ) filtered noise H ( f ) = Ht ( f )H c ( f )H r ( f ) Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 132 : “shiv rpi” .

Nyquist bandwidth constraint Nyquist bandwidth constraint (on equivalent system): The theoretical minimum required system bandwidth to detect Rs [symbols/s] without ISI is Rs/2 [Hz]. a system with bandwidth W=1/2T=Rs/2 [Hz] can support a maximum transmission rate of 2W=1/T=Rs [symbols/s] without ISI. Showing how efficiently the bandwidth resources are used by signaling techniques. Rs Rs 1 = ≤W ⇒ ≥ 2 [symbol/s/Hz] 2T 2 W Bandwidth efficiency. R/W [bits/s/Hz] : An important measure in DCs representing data throughput per hertz of bandwidth. Equivalently. Shivkumar Kalyanaraman 133 : “shiv rpi” Rensselaer Polytechnic Institute .

Equiv System: Ideal Nyquist pulse (filter) Ideal Nyquist filter H( f ) T Ideal Nyquist pulse h(t ) = sinc(t / T ) 1 −1 2T 0 1 2T f − 2T − T 0 T 2T t W= Rensselaer Polytechnic Institute 1 2T 134 Shivkumar Kalyanaraman : “shiv rpi” .

Example of Nyquist filters: Raised-Cosine filter Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 135 : “shiv rpi” .Nyquist pulses (filters) Nyquist pulses (filters): Pulses (filters) which result in no ISI at the sampling time. Nyquist filter: Its transfer function in frequency domain is obtained by convolving a rectangular function with any real evensymmetric frequency function Nyquist pulse: Its shape can be represented by a sinc(t/T) function multiply by another time function.

Pulse shaping to reduce ISI Goals and trade-off in pulse-shaping Reduce ISI Efficient bandwidth utilization Robustness to timing error (small side lobes) Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 136 : “shiv rpi” .

Raised Cosine Filter: Nyquist Pulse Approximation | H ( f ) |=| H RC ( f ) | 1 r =0 r = 0.5 r =1 r = 0.5 r =1 0. 5 h(t ) = hRC (t ) 1 0.5 r =0 −1 − 3 −1 T 4T 2T 0 1 3 2T 4T 1 T − 3T − 2T − T 0 T 2T 3T Rs Baseband W sSB= (1 + r ) 2 Rensselaer Polytechnic Institute Passband W DSB= (1 + r ) Rs Shivkumar Kalyanaraman 137 : “shiv rpi” .

Raised Cosine Filter Raised-Cosine Filter A Nyquist pulse (No ISI at the sampling time) for | f |< 2W0 − W ⎧1 ⎪ ⎪ 2 ⎡ π | f | +W − 2W0 ⎤ H ( f ) = ⎨cos ⎢ ⎥ for 2W0 − W <| f |< W W − W0 ⎣4 ⎦ ⎪ ⎪0 for | f |> W ⎩ cos[2π (W − W0 )t ] h(t ) = 2W0 (sinc(2W0t )) 1 − [4(W − W0 )t ]2 Excess bandwidth: W − W 0 Rensselaer Polytechnic Institute W − W0 Roll-off factor r = W0 0 ≤ r ≤1 Shivkumar Kalyanaraman 138 : “shiv rpi” .

Pulse Shaping and Equalization Principles No ISI at the sampling time H RC ( f ) = H t ( f ) H c ( f ) H r ( f ) H e ( f ) Square-Root Raised Cosine (SRRC) filter and Equalizer H RC ( f ) = H t ( f ) H r ( f ) H r ( f ) = H t ( f ) = H RC ( f ) = H SRRC ( f ) 1 He ( f ) = Hc ( f ) Taking care of ISI caused by tr. filter Taking care of ISI caused by channel Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 139 : “shiv rpi” .

Pulse Shaping & Orthogonal Bases Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 140 : “shiv rpi” .

Virtue of pulse shaping Rensselaer Polytechnic Institute Source: Rappaport book. chap 6 PSD of a BPSK signal Shivkumar Kalyanaraman 141 : “shiv rpi” .

Waveform Third pulse t/T First pulse Second pulse Data symbol Shivkumar Kalyanaraman 142 : “shiv rpi” Rensselaer Polytechnic Institute . [V] Square-root Raised-Cosine (SRRC) pulse shaping Baseband tr.Example of pulse shaping Amp.

**Example of pulse shaping …
**

Raised Cosine pulse at the output of matched filter

Amp. [V]

Baseband received waveform at the matched filter output (zero ISI)

t/T

Rensselaer Polytechnic Institute

**Shivkumar Kalyanaraman 143
**

: “shiv rpi”

Eye pattern

Eye pattern:Display on an oscilloscope which sweeps the system

response to a baseband signal at the rate 1/T (T symbol duration) Distortion due to ISI amplitude scale Noise margin

**Sensitivity to timing error Timing jitter
**

Rensselaer Polytechnic Institute

**time scale Shivkumar Kalyanaraman
**

144

: “shiv rpi”

**Example of eye pattern:
**

Binary-PAM, SRRC pulse

Perfect channel (no noise and no ISI)

Rensselaer Polytechnic Institute

**Shivkumar Kalyanaraman 145
**

: “shiv rpi”

SRRC pulse … AWGN (Eb/N0=20 dB) and no ISI Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 146 : “shiv rpi” .Example of eye pattern: Binary-PAM.

Example of eye pattern: Binary-PAM. SRRC pulse … AWGN (Eb/N0=10 dB) and no ISI Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 147 : “shiv rpi” .

Bounds Modulation Schemes. Performance. Constellations Pulse Shaping Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 148 : “shiv rpi” .Summary Digital Basics Modulation & Detection.

Extra Slides Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 149 : “shiv rpi” .

Q Representation Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 150 : “shiv rpi” .Bandpass Modulation: I.

AM: linear dependence on quality & power of rcvd signal Spectrally efficient but susceptible to noise & fading Fading improvement using in-band pilot tones & adapt receiver gain to compensate Non-constant envelope: Power inefficient (30-40%) Class A or AB power amps needed: ½ the talk time as FM! Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 151 : “shiv rpi” .Analog: Frequency Modulation (FM) vs Amplitude Modulation (AM) FM: all information in the phase or frequency of carrier Non-linear or rapid improvement in reception quality beyond a minimum received signal threshold: “capture” effect. Better noise immunity & resistance to fading Tradeoff bandwidth (modulation index) for improved SNR: 6dB gain for 2x bandwidth Constant envelope signal: efficient (70%) class C power amps ok.

Example Analog: Amplitude Modulation Rensselaer Polytechnic Institute Shivkumar Kalyanaraman 152 : “shiv rpi” .

- receiver by purushottam.ppt
- STEL-2000A Data Sheet
- dcbyDevilsDuke
- DigitalCommthry Compiled Kumarswamy
- Definitions Dcom
- ECE - Digital Communication
- 5980-1303E-Design of CDMA 2000 Network
- Digital Communications
- 9783642127427-t1 (1).pdf
- Comparison of Performance Metrics for QPSK and OQPSK Transmission Using Root Raised Cosine & Raised Cosine Pulse-shaping Filters for Applications in Mobile Communication
- Digital Comms -Tutorial- Parte I
- 2013_V5_BPSKpiAWGN
- BPSK-SUPERREGEN
- DigiComms
- 2MARKS Q&A
- Report on Digital Signal(Bpsk,Qpsk) Generation,Transmission & Reception (200)
- Electronics & Communication Engineering
- Latex File
- Analysis of Performance Improvement Using GMSK Modulation Technique in Wireless DS-CDMA Communication Systems
- Q
- r05310404 Digital Communications
- QPSK system
- 08r05310404 Digital Communications
- QPSK Description
- Digital Communications Qb Pravindra
- Cognitive Radios
- Spectrum Sensing and Blind Automatic Modulation Classification in Real Time
- Digital Modulations Lecture
- 4.modn
- CMPE457-Chapter 5 - Analog Transmission
- wbn2007-pt-pt-I

Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

We've moved you to where you read on your other device.

Get the full title to continue

Get the full title to continue reading from where you left off, or restart the preview.

scribd