Wireless Communication System

Transcription

1 Wireless Communication System Generic Block Diagram An t PC An r Source Tx Rx Destination P t G t L p G r P r Source a source of information to be transmitted Destination a destination of the transmitted information Tx and Rx transmitter and receiver Ant & Anr Tx and Rx antennas PC = propagation channel Tx includes coding/modulation circuitry (or DSP), power amplifiers, frequency synthesizers etc. Rx includes LNA, down conversion, demodulation, decoding etc. Examples: cellular phones, radio and TV broadcasting, GPS, cordless phones, radar, etc. Main advantages: flexible (service almost everywhere), low deployment cost (compare with cable systems). Main disadvantages: PC is very bad, limits performance significantly, almost all development in wireless com. during last 50 years were directed to combat PC. Lecture 2 18-Sep-17 1(27)

2 Radio Transmitter carrier modulated signal An. LO Mod. PA message Local oscillator (LO) generates the carrier Modulator (Mod.) modulates the carrier using the message signal Power amplifier (PA) amplifies the modulated signal to required power level Antenna (An.) radiates the modulated signal as an electromagnetic wave Check-up question: why modulation? An example of modulation: DSB-SC AM, ( ) = ( ) cos( 2π ) x t m t A f t For more info and refreshment, see your ELG3175 (or ELG4176) textbook. c c Lecture 2 18-Sep-17 2(27)

3 Lecture 2 18-Sep-17 3(27)

4 Wireless Propagation Channel An t PC An r Source Tx Rx Destination P t G t L p G r P r This is a major obstacle to reliable and high quality wireless communications. Why? (3 key reasons) Large signal attenuation. High degree of variability (in time, space, etc.) Out of designer s control (almost completely). Difference between wired (AWGN) and wireless (e.g., Rayleigh) channels. Much effort is spent on modeling, characterization and simulations of the wireless propagation channel. Classification of Channel Models System level / propagation (electromagnetic) based Deterministic / stochastic Theoretical / empirical (semi-empirical) Various techniques, goals, accuracies. Lecture 2 18-Sep-17 4(27)

5 Example of a system-level channel model: In: x(t) g(t) Noise: ξ( t) + Out: y(t) PC y( t) = g( t)* x( t) + ξ ( t) g is the channel gain; may be an impulse response, g( t ), or a frequency response, g( f ). Channel can be LTI, but it can also be time-varying. Almost all channels are modeled as linear. Example of propagation-based model: free space model (or 2-ray model), to be discussed later on. All the models above are deterministic. Example of a stochastic model: Rayleigh channel. All the models above are theoretical. Example of an empirical model: Okumura-Hata model. Lecture 2 18-Sep-17 5(27)

6 Threshold Effect All the communication systems exhibit a threshold effect: when signal-to-noise (SNR) ratio drops below a certain value (called threshold value), the system either doesn t operate at all, or operates with unacceptable quality. The SNR is signal power P γ = SNR = = s (2.1) noise power P n For acceptable performance, where γ γ P P (2.2) th r th γ = SNR (at the Rx), γ th = a threshold SNR, P r = the Rx power, P = the threshold Rx power (sensitivity), th i.e. Pr is bounded from below to provide satisfactory performance. Outage event: if γ < γth Pr < Pth unsatisfactory system performance). It is very important to evaluate correctly the Rx signal power r P when designing a communication system (link). Lecture 2 18-Sep-17 6(27)

7 P th is affected by: ELG4179: Wireless Communication Fundamentals S.Loyka 1) Bandwidth 2) Rx noise figure 3) Type of modulation (i.e, BPSK/QPSK/16-QAM); 4) Coding (i.e, no coding, (7, 4) Hamming code, etc.) Noise power: In an additive white (thermal) noise channel, the SNR can be expressed as γ = P / P (2.3) where P0 = kt ff = Rx noise power, 23 k = J / K = Boltzman constant, T = Rx temperature ( 0 K), f = equivalent (noise) bandwidth, F = Rx noise figure (typically a few dbs). γ th is found based on a desirable error rate performance. r 0 Lecture 2 18-Sep-17 7(27)

8 Link Budget Analysis The link budget relates the Tx power, the Rx power, the path loss, Rx noise and additional losses and margins into a single equation: where P r G G = (2.4) P T R t L P L a F s P t = Tx power, G T = Tx antenna gain, G R = Rx antenna gain, L P = Propagation path loss, L a = Additional losses (i.e., cable, aging, etc.) F = Fading (and other) margin; s Interference margin: can be added when the system operates in interference environment (i.e, cellular). Fading margin: can be added when the system operates in a fading channel. F s = 1 if no fading. Link margin = threshold. γ / γ th how far away the link is from the Link budget equation (2.3) can be used to find the required Tx power (P t ) or the maximum acceptable path loss (L P ). This is a first step in the design of a wireless system (link). Lecture 2 18-Sep-17 8(27)

9 Given P r ELG4179: Wireless Communication Fundamentals S.Loyka An example Lp = 10 W( 90 dbm); = 150 db (10 ); G = 10 db; G = 10 db; L = 0 db; r t a find required P t : Pr Lp Pt = ( Pt = Pr + Lp Gt Gr )[ db] = 40dBm 10 W; G G t r Minimum P r ( sensitivity ) for a WiFi router Rx: db, dbm, etc.: power ratio: dB = 10lg 10(100) amplitude (voltage/current) ratio: dB = 20lg 10(100) dbm = db w.r.t. 1mW: 20dBm 1mW 1020/10 = 100mW = 0.1W Lecture 2 18-Sep-17 9(27)

10 Effect of Interference No interference: P γ = SNR = s γ th (2.6) P n As a simple model, assume that interference acts like noise: P SNR γ = SNIR = s = γ P + P 1+ INR n i th (2.7) where INR= P / i P = interference-to-noise ratio (INR). n Satisfactory performance typically requires γ ~ 10dB. Minimum required received power under no interference: Pr = Ps γ thpn (2.8) Minimum received power with interference: th ( ) ( 1 INR ) P γ P + P = γ P + (2.9) r th n i th n Compare (2.9) to (2.8): the effect of interference is to boost the required Rx and thus Tx powers. Noise dominated systems vs. interference dominated systems. Lecture 2 18-Sep-17 10(27)

11 Three Factors in the Propagation Path Loss: The propagation path loss is where L A = average path loss, L LF = large-scale fading, L = small-scale fading. SF LP = LALLF LSF (2.5) Propagation Path Loss Components Siwiak, Radiowave Propagation and Antennas for Personal Communications, Artech House, 1998 Lecture 2 18-Sep-17 11(27)

12 Propagation Channel: Basic Mechanisms Approximations are very important! LOS propagation: consider a communication link in free space An t PC An r Tx Rx P t G t L p G r P r Assume for a moment that the Tx antenna is isotropic, then power flux density at distance R is P Π t i = (2.10) 2 4 π R Since the antenna is not isotropic, PG Π = t t (2.11) 2 4 π R Equivalent isotropic radiated power (EIRP) is P = PG (2.12) e t t This is the power radiated by isotropic antenna, which produces the same power at the receiver as our non-isotropic antenna. Lecture 2 18-Sep-17 12(27)

13 Effective Aperture & Received Power: Free Space Effective aperture of Rx antenna, Se: G R 4π = S S λ 2 e e 2 λ = GR (2.13) 4 π Power received by Rx antenna is λ G P rpe r = Π Se = GtGr Pt = ; 4πR L where Lp is the propagation loss (Friis equation), L p 4πR = λ 2 2 p (2.14) (2.15) Friis equation (2.15) is valid in the far field only: 2 2D R & R D, λ (2.16) λ where D is the maximum antenna size. Usually R > λ and 2 nd part reduces to R D; 1 st part dominates in many cases. Free space propagation model is simple, but unrealistic. Real environments are more complex. However, the free space model provides good starting point for more complex models. Lecture 2 18-Sep-17 13(27)

14 Relation between the power flux density Π and electric field magnitude E : 2 E Π =, W0 = 120 π [ Ω] 377 Ω W 0 where W0 = E / H is the free space wave impedance. Wavelength and frequency are related: c λ = ct = λ [m] = f 300 f [MHz] where c=3*108 [m/s] speed of light, T=1/f the period. Another form of the Friis equation: L p 2 2 4πR 4πRf = = λ c Q.: For given Pt, G t, and d, show that E at distance d can be expressed, in free space, as E = 30PG t t d Lecture 2 18-Sep-17 14(27)

15 Path Loss: Wireless (LOS) vs. Cable 0 20 Path gain (db) vs. distance (m) LOS, 1 GHz LOS, 10 GHz Cable, 0.1dB/m Cable, 1dB/m Lecture 2 18-Sep-17 15(27)

16 Three Basic Propagation Mechanisms Reflection: EM wave impinges on an object of very large size (much greater than λ), like surface of Earth; large buildings, mountains, etc. Diffraction: the Tx-Rx path is obstructed by an object or large size (>> λ ), maybe with sharp irregularities (i.e. edges). Secondary waves are generated (i.e. bending of waves around the obstacle). Scattering: the medium includes objects or irregularities of small size (<<λ). Examples: rough surface, rain drops, foliage, atmospheric irregularities (>10GHz). Diffraction: direction of propagation differs from ray optics predictions. All three mechanisms are important in general. Individual contributions vary on case by case basis. In order to model accurately the PC, one must be able to model all 3 mechanisms. Lecture 2 18-Sep-17 16(27)

17 Propagation Mechanisms: Illustrations LOS Ground station Ground multipath Reflection LOS Ground station Ground multipath Scattering (diffuse reflection) Scattering & Reflection: specular and diffuse For more information, see S. Loyka, A. Kouki, Using Two Ray Multipath Model for Microwave Link Budget Analysis, IEEE AP Magazine, v. 43, N. 5, pp , Oct Lecture 2 18-Sep-17 17(27)

18 P.M. Shankar, Introduction to Wireless Systems, Wiley, Lecture 2 18-Sep-17 18(27)

19 Propagation Loss Components In terms of signal variation in space (i.e. distance) and time, there are 3 main factors as well, in propagation path loss: Attenuation: average signal power vs. distance ignoring small and large-scale variations; keep only very large-scale effects, i.e. spreading of power with distance as in free space. Large-scale fading (shadowing): over ~ 100m, ignoring variations over a few wavelengths and smaller. Small-scale fading (multipath): over fraction of λ to few λ. Siwiak, Radiowave Propagation and Antennas for Personal Communications, Artech House, 1998 Lecture 2 18-Sep-17 19(27)

20 Average path loss (attenuation) similar to free space, but path loss exponent may be different. The average received power Pra is P 1 r = PG t T G aν R ~ ; R ν R ν ν = (2.17) In free space, ν = 2; in general, it depends on environment; in practice, it is obtained from measurements ( ν = ). Smart antennas are useful in combating all three factors, but they are most efficient for #3 (small-scale fading). While the average path loss is modeled deterministically, large and small scale fading are modeled as random variables (processes). The received power under large and small scale fading is: P = g g P (2.17a) r l s ra where g l and g s are the large and small scale fading factors (typically modeled as log-normal and Rayleigh random variables). Lecture 2 18-Sep-17 20(27)

21 Two-Ray (ground reflection) Model Total received field is E = E + E e E E t D R D t j ϕ A A Γ = ; ER = d d + d D A d = 1+ Γ D d d + d D e j ϕ E D - direct (line-of-sight LOS) component, E R - reflected component, ϕ - phase difference Γ - complex reflection coefficient (2.18) Lecture 2 18-Sep-17 21(27)

22 Ground reflection affects the path loss significantly. The phase difference is In many cases, 2π 2π ϕ = ( d1 + d2 dd) = d λ λ (2.19) d d1 + d2, d ; 1, 2,, and D D λ d d dd ht hr 1 d + d (2.20) For small α ( α 1) Γ Under these approximations, the total received field becomes: A 4πh h A 20h h Et 1 e, R > d λr λ Note that total receive power D The path loss is j ϕ t r t r 2 L p P r t r Compare with free space (FS): ~ Et ~ R (2.21) R = (2.22) h h 1 P r, FS ~ 2 R, 4πR LFS = λ Conclusion: multipath can significantly affect the path loss! Composite model in practice: L = max{ L, L,1} 2 ray FS 2 Lecture 2 18-Sep-17 22(27)

23 Example of Two-Ray Path Loss Path gain (db) vs. distance (lambda) ε = 16, h = h = 2 r t r 20ht hr 80 λ 4ht hr 16 λ P r ~ 1 2 R P r ~ 1 4 R 140 near field far field vertical horizontal conductor free space two ray approx. Q.: do it yourself using Matlab. Lecture 2 18-Sep-17 23(27)

24 Propagation of Electromagnetic (Radio) Waves Digital and Analog Communication Systems, Eighth Edition by Leon W. Couch II Lecture 2 18-Sep-17 24(27)

25 LOS and Radio Horizon Two-ray model is valid as long as there is an LOS Tx-Rx path, ( ) R < d 4 h + h [km] LOS t r where ht, h r are in meters, and d LOS is the maximum LOS distance in km. This is so-called radio-horizon. When R > dlos, the LOS as well as reflected paths are obstructed by Earth and path loss increases significantly due to extra diffraction loss. Two-ray model cannot be used. Lecture 2 18-Sep-17 25(27)

26 Rough Surface: Scattering Rough surface -> Rayleigh criterion: λ h (2.23) 8sinα where h is r.m.s. variation in surface height Flat surface reflection coefficient is multiplied by a scattering loss factor: where h 0 /( sin ) 2 2 h σ I h 0 0 h0 σ ρ s = exp 8 8 h (2.24) = λ π α, σ h is standard deviation of the surface height, I 0 is the modified Bessel function of 1st kind and zero order Modified reflection coefficient: Γ = Γ ρ (2.25) s Lecture 2 18-Sep-17 26(27)

27 Summary Wireless propagation channel Various types of channel models Link budget analysis; effect of interference Three propagation mechanisms Path loss exponent Free-space propagation Ground reflection and two-ray model Rough surface and scattering o Rappaport, Ch. 4. Reading: References: o S. Salous, Radio Propagation Measurement and Channel Modelling, Wiley, (available online) o J.S. Seybold, Introduction to RF propagation, Wiley, o Other books (see the reference list). Note: Do not forget to do end-of-chapter problems. Remember the learning efficiency pyramid! Lecture 2 18-Sep-17 27(27)