The Abdus Salam International Centre for Theoretical Physics ICTP-ITU/BDT School on Sustainable Wireless ICT Solutions for Environmental Monitoring

6 – 22 February 2012, ICTP, Miramare-Trieste, Italy

RF Channel & Modulation Fundamentals

Prof. Ryszard Struzak National Institute of Telecommunications, Poland r.struzakATieee.org Objective

• Review of basic physical processes involved in modern communication systems that may be needed to understand better the further training/ teaching activities at ICTP

(CC) R Struzak 2 Topics for discussion

• Radio transmission • Modulation schemes • Constellation diagrams • Modulation spectra

(CC) R Struzak 3 http://ursi.org/files/RSBissues/ RSB_339_2011_12.pdf

(CC) R Struzak 4 (CC) R Struzak 5 Source: URSI RSB No 339 Dec. 2011 (CC) R Struzak 6 Source: URSI RSB No 339 Dec. 2011

(CC) R Struzak 7 Radio Transmission System

• A message, from a source of messages, is delivered to a distant destination via radio communication channel using energy of radio waves » Message in its most general meaning is the object of communication. Depending on the context, the term may apply to both the information contents and its actual presentation, or signal. » The baseband signal usually consist of a finite set of symbols. E.g. text message is composed of words that belong to a finite vocabulary of the language used. Each word in turn is composed by letters of a (finite) alphabet. (Analog-to-digital conversion) » The channel consists of a transmitter node, radio propagation path and receiver node. • The radio propagation process distorts the signal » Predictable vs unpredictable distortions • The transmitter & receiver pair process the signal to repair the signal distortions following common communication protocol and communication policy

(CC) R Struzak 8 Volume of information transmitted by isolated communication channel

C = B*log2{1 + [S/(No*B)]}

Noise power density, W/Hz Received signal power, W Bandwidth, Hz Capacity (maximum throughput), bit-per-second • The capacity to transfer error-free information is enhanced with increased bandwidth B, even though the signal-to-noise ratio is decreased because of the increased bandwidth. • C/B (bps/Hz) → bandwidth efficiency

(CC) R Struzak 9 • The actual throughput can only approach the channel capacity using an extremely complicated encoder. – The statistical properties of the signal approach those of Gaussian noise and the coding time increases indefinitely. – Rice (1950) has shown that in order to achieve a data rate which is 96 percent of the channel capacity with S/NoB = 10 and an error probability of 10-5 the number of block encoded channel signals required is 210,000. This is just as impractical as theinfinite coding delay required. • Source: R. F. Linfield: Radio Channel Capacity Limitations; ITS NTIA

(CC) R Struzak 10 Frequency V division Frequency slotB Time slot T

V = B*T*log2{1 + [S/(No*B)]}

Time division headers; interactive & real-time applications that require acknowledgment!

(CC) R Struzak 11 FDD radio links

• Two stations can talk and listen to each other at the same time (time-sharing). • This requires separate synchronized frequency channels – a technique known as Frequency Division Duplex (FDD)

Frequency Channel 2 TX RX Station 1 Station 2 RX TX Frequency Channel 1

(CC) R Struzak 12 TDD radio links

• Two stations can talk and listen to each other using the same frequency channel (frequency-sharing), one after another. • This requires time synchronization/ handshaking – a technique known as Time Division Duplex (TDD)

Single Frequency Channel TX TX Station 1 Station 2

RX RX

(CC) R Struzak 13 Broadband (INTEL: 100+ Mbps)

INTEL opinion: 100 Mbps necessary! Ch. Legutko (Intel Corp. Wireless Standards & Regulations Mngr): „Regulatory & Spectrum”; presentation, Broadband Forum, Warsaw, July 2006 (CC) R Struzak 14 !"#$%&'()*+)+#,-*.()&/$/-/01),*$$20/,#-/*0),3#00'&),#%#,/-4))

534(/,#&)) 6',30*&*1/,#&) 7'12&#-*.4) !,*0*$/,)))))) ,*0(-.#/0-() ,*0(-.#/0-() ,*0(-.#/0-() ,*0(-.#/0-()

8#09:/9-3) =%2./*2() A.'B2'0,4) C*(-)) '$/((/*0)) #((/10$'0-() ;*/(') >).','%-/*0) 5*:'.) !0

(CC) R Struzak 15 • The fundamental physical limitations are bandwidth and noise. • The bandwidth limits the number of signaling elements or symbols which can be transmitted per unit time. • The noise limits the number of distinguishable features on each symbol. • The combined effects limit the capacity of a communication system.

(CC) R Struzak 16 Radio Channel Model Original message/ data Coding/ Transmitter Processing Time series T-antenna

EM waves: time- Propagation medium distance- direction- polarization

Environment R-antenna Noise, Fading, Time series Receiver Processing/ De-coding Reconstructed message/ data

(CC) R Struzak 17 SISO, SIMO, MISO, MIMO

1

SISO Tx Rx MISO Tx Rx m

1 1 1

SIMO Tx Rx MIMO Tx Rx

n m n

Oestges C, Clerckx B, "MIMO Wireless Communications : From Real-world Propagation to Space-time Code Design," Academic, 2007 http://en.wikipedia.org/wiki/MIMO

(CC) R Struzak 18 Radio transmitter

1. Generates a RF carrier 2. Combines it with the baseband signal into a RF signal through modulation • Performs additional operations E.g. analog-to-digital conversion, formatting, coding, spreading, adding additional messages/ characteristics such as error-control, authentication, or location information 3. Radiates the resultant signal in the form of modulated radio wave - it maps the original message into the radio-wave signal launched at the transmitting antenna

(CC) R Struzak 19 Propagation process

• Transforms (maps) the radio-wave launched by the transmitter into the incident radio wave at the receiver antenna • Depends on extra variables (e.g. distance), additional radio waves (e.g. reflected wave, environmental waves), random uncertainty additive (noise), multiplicative (fading), etc. • A separate talk

(CC) R Struzak 20 Radio receiver:

1. Extracts the wanted signal from the Wanted signal incident radio waves (a solid “window” in the signal hyperspace) Noise 2. Recovers the original message through • reversing the transmitter operations (demodulation, decoding, de-spreading, etc.), + Receiver • compensating propagation transformations, and • correcting transmission distortions

• Maps the incident signals into the Unwanted signals recovered message

• Coexistence - Separate talk! Unwanted signals and nonlinearities are often ignored! (CC) R Struzak 21 Modulation & Demodulation

• Modulation - translation the baseband message signal to radio frequencies • Demodulation -- the reverse process) – Single carrier: • continuous (sinusoidal), • Walsh functions (http://mathworld.wolfram.com/WalshFunction.html)

• Random carriers (http://en.wikipedia.org/wiki/Category:Radio_modulation_modes) – Multiple carriers (e.g. OFDM)

(CC) R Struzak 22 Modulation Process

ffaaaat= ( 123, , ,...n ,) (= carrier)

aaa123, , ,... an (= modulation parameters) t (= time) • Moves information from baseband to RF • Implies varying one or more characteristics

(modulation parameters a1, a2, … an) of a carrier f in accordance with the information-bearing (modulating) baseband signal.

(CC) R Struzak 23 Frequency translation

Signal's baseband bandwidth is its bandwidth before modulation and multiplexing, or after demultiplexing and demodulation

• Signal "at baseband" comprises all frequency components carrying information. Modulation shifts the signal up to RF frequencies to allow for radio transmission. Usually, the process increases the signal bandwidth. Steps are often taken to reduce this effect, such as filtering the RF signal prior to transmission.

(CC) R Struzak 24 0 802.11g

-10

-20 dB -30

-40

-50 -50 -40 -30 -20 -10 0 10 20 30 40 50 Frequency from centre, MHz

(CC) R Struzak 25 802.11b/g channels (Center frequencies in GHz)

1. 2,412 8. 2,447 • Different subsets of these channels are 2. 2,417 9. 2,452 made available in 3. 2,422 10. 2,457 various countries 4. 2,427 11. 2,462 • Spacing: ~5 (12) MHz 5. 2,432 12. 2,467 • Occupied bandwidth: 6. 2,437 13. 2,472 ~22 MHz (802.11b), ~16.6 MHz (802.11g) 7. 2,442 14. 2,484

Source:(CC) R Morrow: R Struzak Wireless network coexistence, p. 198 26 Why translation?

• Effective radiation of EM energy requires antenna dimensions to be comparable with the wavelength: – 3 kHz: wavelength 100 km – 3 GHz: wavelength 10 cm • Allows shared use of the radio transmission medium

(CC) R Struzak 27 Continuous sinusoidal carrier

A sin[ωt +ϕ]; A, ω, or/and ϕ = var. • Amplitude • Frequency Phase modulation modulation modulation (AM) (FM) (PM) – A = A(t) – A = const – A = const – ω = const – ω = ω(t) – ω = const – ϕ = const – ϕ = const – ϕ = ϕ(t)

(CC) R Struzak 28 AM, FM, PM

The modulating signal superimposed on the carrier wave & the resulting modulated signal. [wikipedia]

Java simulation: http://www.home.agilent.com/upload/cmc_upload/All/ Amplitude_Modulation.htm?cmpid=zzfindnw_amplitudemod

(CC) R Struzak 29 Modulation Spectra

• The Nyquist bandwidth is the minimum bandwidth

that can represent a Nyquist Minimum Adjacent signal (within an Bandwidth Channel acceptable error) • The spectrum occupied by a signal should be as close as practicable to that minimum, otherwise adjacent channel interference occur • The spectrum occupied

Relative Magnitude (dB) Magnitude Relative by a signal can be reduced by application of filters

Frequency

(CC) R Struzak 30 Digital modulation

• Modulation that uses a finite number of distinct states (signals) to represent data • Classic types: – ASK (Amplitude-shift keying): uses a finite number of amplitudes; – FSK (Frequency-shift keying): uses a finite number of frequencies – PSK (Phase-shift keying):uses a finite number of phases – QAM (Quadrature amplitude modulation): uses at least 2 phases, and at least 2 amplitudes » PSK can be regarded as a special case of QAM (magnitude of the signal is a constant and the phase is varying. This can also be extended to FSK as a special case of phase modulation. • Other: » Multiple-carrier modulation, » Carrierr-less modulation, » Walsh-type modulation » Differential modulation • Combinations of the above.

(CC) R Struzak 31 Effect of pulse width on spectral content. The shorter pulse width the wider the spectrum

(CC) R Struzak 32 33 (CC) R Struzak Differential Modulation • At transmitter: each symbol is modulated relative to the previous symbol and modulating data • DPSK (Differential Phase-Shift Keying) example: 0 = no change; 1 = change phase by +1800 • At receiver: the current symbol is reproduced using the previous symbol as reference. • The previous symbol serves as an estimate of the channel. A no-change condition causes the modulated signal to remain at the same 0 or 1 state as the previous symbol. • Differential modulation has 2 sources of error: • a corrupted current symbol, and corrupted reference (the previous symbol)

(CC) R Struzak 34 ASK: Amplitude Shift Keying

Baseband Data 1 0 0 1 0 ASK (binary) signal Acos(ωt) Acos(ωt)

• Pulse shaping can be employed to remove spectral spreading • ASK demonstrates poor performance, as it is heavily affected by noise, fading, and interference during propagation • Can be expanded to M-ary scheme, with multiple amplitudes

(CC) R Struzak 35 FSK: Frequency Shift Keying

Baseband Data 1 0 0 1 BFSK (Binary Frequency Shift Keying) signal f1 f0 f0 f1

s0 =Acos(ωc-Δω)t; s1 =Acos(ωc+Δω)t

• Example: The ITU-T V.21 modem standard uses FSK • BFSK can be expanded to M-ary scheme, with multiple frequencies

(CC) R Struzak 36 PSK: Phase Shift Keying

Baseband Data 1 0 0 1 BPSK (Binary Phase Shift Keying)

signal s1 s0 s0 s1

s0 =-Acos(ωct); s1 =Acos(ωct) • BPSK – better performance than ASK and BFSK • Drawback: rapid amplitude change between symbols due to phase discontinuity requires infinite bandwidth. • BPSK can be expanded to M-ary scheme, with multiple phases

(CC) R Struzak 37 • Phase Shift Keying (PSK) is often used as it provides efficient use of RF spectrum. π/4 QPSK (Quadrature PSK) reduces the envelope variation of the signal. • High level M-array schemes (such as 64-QAM) are bandwidth-efficient but susceptible to noise and require linear amplification • Constant envelope schemes (e.g. GMSK) allow for non-linear power-efficient amplifiers • Coherent reception provides better performance but requires a more complex receiver

(CC) R Struzak 38 Constellation diagram

= graphical representation of the amplitude and phase of each possible symbol state in time domain – The x-axis represents the in-phase component and the y-axis the quadrature component of the complex envelope – The distance between signal states on a constellation diagram indicates • how different the modulation waveforms are • how easily a receiver can differentiate between them.

(CC) R Struzak 39 Example: BPSK Constellation Diagram

Q

I o o (|1|, -180 ) (|1|, 0 )

(CC) R Struzak 40 QPSK

• Quadrature Phase Shift Keying (QPSK) can be interpreted as two independent BPSK systems (one on the I-channel and one on Q), and thus the same performance but twice the bandwidth efficiency • Large envelope variations occur due to abrupt phase transitions, thus requiring linear amplification

(CC) R Struzak 41 QPSK Constellation Diagram

Q Q

I I

Carrier phases Carrier phases {0, /2, , 3 /2} π π π {π/4, 3π/4, 5π/4, 7π/4}

• Quadrature Phase Shift Keying has twice the bandwidth efficiency of BPSK since 2 bits are transmitted in a single modulation symbol

(CC) R Struzak 42 QPSK examples

Q Q Q

I I I

Conventional QPSK Offset QPSK π/4 QPSK • All QPSK schemes require linear power amplifiers • Conventional QPSK has transitions through zero (i.e. 1800 phase transition) which requires highly linear amplifiers. • In Offset QPSK, the phase transitions are limited to 900, the transitions transitions through zero do not occur. • In π/4 QPSK transitions through zero do not occur. This scheme produces the lowest envelope variations.

(CC) R Struzak 43 Multi-level (M-ary) Phase and Amplitude Modulation

16 QAM 16 PSK 16 APSK

• Amplitude and phase shift keying can be combined to transmit several bits per symbol. – Often referred to as linear as they require linear amplification. – More bandwidth-efficient, but more susceptible to noise. • For M=4, 16QAM has the largest distance between points, but requires very linear amplification. 16PSK has less stringent linearity requirements, but has less spacing between constellation points, and is therefore more affected by noise.

(CC) R Struzak 44 • Animations:

• http://www.home.agilent.com/upload/cmc_upload/All/ IQ_Modulation.htm?cmpid=zzfindnw_iqmod

• http://www.home.agilent.com/agilent/editorial.jspx? cc=PL&lc=eng&ckey=1756523&nid=-34944.0&id=1756523

(CC) R Struzak 45 Comparison

Modulation C/B (bandwidth „Error-free” Eb/No Format efficiency) 16 PSK 4 18 dB 16 QAM 4 15 dB 8 PSK 3 14.5 dB 4 PSK 2 10 dB 4 QAM 2 10 dB BFSK 1 13 dB BPSK 1 10.5 dB

(CC) R Struzak 46 Distortions

Perfect channel White noise Phase jitter

(CC) R Struzak 47 Carrier-less or Pulse-Carrier

• Carrier: A train of identical pulses regularly spaced in time

(CC) R Struzak 48 (CC) R Struzak 49 Ultra-Wideband (UWB) Systems

• Radio systems that use time-domain processing (e.g., pulse-position modulation) for communications, or sensing applications. – e.g. short-range radar; – Also called carrier-less or impulse systems; – Occupy frequency bandwidth >25% of the center frequency or >1.5 GHz. – Use narrow pulses • Of the order of 1 to 10 nanoseconds.

(CC) R Struzak 50 Multiple-carrier systems example

• OFDM = Orthogonal Frequency Division Multiplex • Basic idea: – Using a large number of parallel narrow-band sub-carriers instead of a single wide-band carrier to transport information • Advantages – Efficient in dealing with multi-path and selective fading – Robust again narrow-band interference • Disadvantages – Sensitive to frequency offset and phase noise – Peak-to-average problem reduces the power efficiency of RF amplifier at the transmitter • Adopted for various standards • DSL, 802.11a, DAB, DVB

(CC) R Struzak 51 OFDM Frequency domain

N carriers

B B

– Data are transmited using only one carrier – Data are shared among several carriers and simultaneously transmitted

– Selective Fading – Flat Fading per carrier – Very short pulses – N long pulses, each pulse length ~1/B of the length ~ N/B

(CC) R Struzak 52 OFDM 2

Inverse Parallel Serial to Fourier Fourier To Serial Transmission Parallel Transform Transform converter converter FFT IFFT

Data Data in Transmit Decode coded time time-domain each in domain samples of frequency frequency one one symbol bin domain symbol independently one at a symbol time at a time

More information: http://www.iss.rwth-aachen.de/Projekte/Theo/OFDM/node2.html

(CC) R Struzak 53 SS communications basics

Original information Spreading Original signal Spread signal

Propagation effects Transmission Unwanted signals + Noise

Reconstructed De-spreading information Spread signal+ Reconstr. signal

(CC) R Struzak 54 Example: GPS

• RF bandwidth Bc ~ 2MHz Filter bw. Bm ~100 Hz

• Processing gain: (2Mhz) / (100Hz) = = 20’000 (~+43 dB)

• Input S/N = -20 dB (signal = 1% of noise [W]) • Output S/N = +23 dB (signal = 200 x noise [W])

(CC) R Struzak 55 SS techniques

• FH: frequency hoping (frequency synthesizer controlled by pseudo-random sequence of numbers)

• DS: direct sequence (pseudo-random sequence of pulses used for spreading)

• TH: time hoping (spreading achieved by randomly spacing transmitted pulses) • Random noise as carrier

• Other techniques (radar and other applications) • Combination of the above

(CC) R Struzak 56 SS: basic characteristics

• Signal spread over a wide bandwidth >> minimum bandwidth necessary to transmit information • Spreading by means of a code independent of the data • Data recovered by de-spreading the signal with a synchronous replica of the reference code – TR: transmitted reference (separate data-channel and reference-channel, correlation detector) – SR: stored reference (independent generation at T & R pseudo-random identical waveforms, synchronization by signal received, correlation detector) – Other (MT: T-signal generated by pulsing a matched filter having long, pseudo-randomly controlled impulse response. Signal detection at R by identical filter & correlation computation) (CC) R Struzak 57 SS Processing Gain =

= [(S/ I)in/ (S/ I)out ] = ~Bc/ Bm

Example: GPS signal RF bandwidth Bc ~ 2MHz Filter bandwidth Bm ~ 100 Hz

Processing gain: (2Mhz)/(100Hz) = 20’000 (~+43 dB)

Input S/N = -20 dB (signal power = 1% of noise power) Output S/N = +23 dB (signal power = 200 x noise power)

(GPS = Global Positioning System) (CC) R Struzak 58 • Beware of misprints!!! These materials are preliminary notes intended for my lectures only and may contain misprints. Feedback is welcome. If you notice serious faults, or you have an improvement suggestion, please describe it exactly and I will try to modify the materials. • This work is licensed under the Creative Commons Attribution License ( http://creativecommons.org/ licenbses/by/1.0) and may be used freely for individual study, research, and education in not-for-profit applications. Any other use requires the written author’s permission. If you cite these materials, please credit the author, title, and place. • These materials and any part of them may not be published, copied to or issued from another Web server without the author's written permission. • Copyright © 2012 Ryszard Struzak.

(CC) R Struzak 59 Thank you

(CC) R Struzak 60