Typical Wireless Communication System

Wireless Communication Fundamentals Part II David Tipper Associate Professor Graduate Telecommunications and Networking Program University it of Pittsburgh Telcom 2700 Slides 3 Typical Wireless Communication System Source Source Encoder Channel Encoder Modulator Channel Destination Source Decoder Channel Decoder Demod -ulator Telcom 2700 2

Modulation and demodulation analog baseband digital signal data digital analog 101101001 modulation modulation radio transmitter radio carrier analog demodulation analog baseband signal synchronization decision digital data 101101001 radio receiver radio carrier Telcom 2700 3 Modulation Modulation Converting digital or analog information to a waveform suitable for transmission over a given medium Involves varying some parameter of a carrier wave (sinusoidal waveform) at a given frequency as a function of the message signal General sinusoid Amplitude A cos (2f C t + ) Frequency Phase If the information is digital changing parameters is called keying (e.g. ASK, PSK, FSK) Telcom 2700 4

Modulation Motivation Smaller antennas (e.g., /4 typical antenna size) = wavelength = c/f, where c = speed of light, f= frequency. 3000Hz baseband signal => 15 mile antenna, 900 MHz => 8 cm Frequency Division Multiplexing provides separation of signals medium characteristics Interference rejection Simplifying circuitry Modulation shifts center frequency of baseband signal up to the radio carrier Basic schemes Amplitude Modulation (AM) Amplitude Shift Keying (ASK) Frequency Modulation (FM) Frequency Shift Keying (FSK) Phase Modulation (PM) Phase Shift Keying (PSK) Telcom 2700 5 Digital Transmission Wireless networks have moved almost entirely to digital modulation Why Digital Wireless? Increase System Capacity (voice compression) more efficient modulation Error control coding, equalizers, etc. => lower power needed Add additional services/features (SMS, caller ID, etc..) Reduce Cost Improve Security (encryption possible) Data service and voice treated same (3G systems) Called digital transmission but actually Analog signal carrying digital data Telcom 2700 7

Digital modulation Techniques Amplitude Shift Keying (ASK): change amplitude with each symbol frequency constant low bandwidth requirements very susceptible to interference Frequency Shift Keying (FSK): change frequency with each symbol needs larger bandwidth Phase Shift Keying (PSK): Change phase with each symbol More complex robust against interference 1 0 1 1 0 1 1 0 1 t t t Telcom 2700 8 Basic Digital Modulation Techniques Telcom 2700 9

Amplitude-Shift Keying One binary digit represented by presence of carrier, at constant amplitude Other binary digit it represented by absence of carrier s t where the carrier signal is Acos(2πf c t) B=2f b, f b = input bit rate Acos 2f t 0 c binary 1 binary 0 Very Susceptible to noise Used to transmit digital data over optical fiber Telcom 2700 10 Binary Frequency-Shift Keying (BFSK) Two binary digits represented by two different frequencies near the carrier frequency s t Acos 1 2f t Acos 2 2f t where f 1 and f 2 are offset from carrier frequency f c by equal but opposite amounts B = 2([f 2 f 1 ]/2 + f b ) Where f b = input bit rate binary 1 binary 0 Telcom 2700 11

Phase-Shift Keying (PSK) Two-level PSK (BPSK) Uses two phases to represent binary digits s t Acos 2f ct A cos 2 f c t Acos 2ff c t Acos2f binary 1 binary 0 cos binary 1 t binary 0 c B = 2f b Telcom 2700 12 Signal Constellation Given any modulation scheme, it is possible to obtain its signal constellation. Represent each possible signal as a vector in a Euclidean space. In symbol detection decode incoming signal as closest symbol in the signal constellation space If we know the signal constellation, we can estimate the performance in terms of the probability of symbol error given the noise parameters. Probability of error depends on the minimum distance between the constellation points. Telcom 2700 13

Advanced Modulation Schemes Variations on ASK, FSK and PSK possible Attempt to improve performance Increase data for a fixed bandwidth Remove requirement for clock recovery Improve BER performance Main schemes for wireless systems are based on FSK and PSK because they are more robust to noise Telcom 2700 14 M-ary Signaling/Modulation What is M-ary signaling? The transmitter considers k bits at a times. It produces one of M signals where M = 2 k. Example: Quarternary PSK (k = 2) Input: Signal : 2E 00 T 01 11 10 2E T 2E T 2E T cos cos cos cos 2f c t, 0 t T 2 f ct, 0 t T 2 2f ct, 0 t T 3 2 f t, 0 t T c Telcom 2700 18 2

QPSK Constellations 2 () (t) 2 () (t) 1 (t) 1 (t) Rotated by /4 Telcom 2700 19 QPSK Telcom 2700 20

M-ary Error Performance A received symbol is decoded into the closest the symbol in the signal constellation As the number of symbols M in the signal space increases the decoding region for each symbol decreases BER goes up For example MPSK, as M increases the bandwidth remains constant, increase in symbol error rate BPSK QPSK Telcom 2700 21 Selection of Encoding/Modulation Schemes Performance in an Noisy channel How does the bit error rate vary with the energy per bit available in the system Performance in fading multipath channels Same as above, but add multipath and fading Bandwidth requirement for a given data rate Also termed spectrum efficiency or bandwidth efficiency How many bits/sec can you squeeze in one Hz of bandwidth Cost The modulation scheme needs to be cost efficient Telcom 2700 23

Performance in Noisy channels AWGN = Additive White Gaussian Noise This has a flat noise spectrum with average noise power of N 0 The probability of bit error (bit error rate) is measured as a function of ratio of the energy per bit E b to the average noise value BER or P e variation with E b /N 0 E b /N 0 is a measure of the power requirements Tradeoffs! Some form of PSK used in most wireless systems P e 10 0 Binary modulation schemes 10-1 10-2 10-3 10-4 10-5 10-6 BPSK DPSK BFSK 2 4 6 8 10 12 14 E b /N 0 Telcom 2700 24 Effect of Mobility? A moving receiver can experience a positive or negative Doppler shift in received signal, depending on direction of movement Results in widening frequency spectrum Fading is a combination of fast fading (Short term fading and Intersymbol interference) and long term fading (path loss- shadowing) [Garg and Wilkes Fig 4.1] Telcom 2700 25

Performance in Fast Fading Channels The BER is now a function of 0 10 the average E b/n 0 The fall in BER is linear 10-1 Large power consumption on average to achieve a good 10-2 BER 30 db is three orders of 10-3 magnitude larger Must use Diversity 10-4 techniques to overcome effect of Short Term (Fast) 10-5 Fading and Multipath Delay P e 10-6 Binary modulation schemes under fading BPSK-No fading BPSK DPSK BFSK 5 10 15 20 25 30 35 40 45 E b /N 0 Telcom 2700 26 What is Diversity? Idea: Send the same information over several uncorrelated forms Not all repetitions will be lost in a fade Types of diversity it Time diversity repeat information in time spaced so as to not simultaneously have fading Error control coding! Frequency diversity repeat information in frequency channels that are spaced apart Frequency hopping spread spectrum Space diversity use multiple antennas spaced sufficiently apart so that the signals arriving at these antennas are not correlated Usually deployed in all base stations but harder at the mobile Telcom 2700 27

Performance Degradation and Diversity Issue Performance Affected Diversity Technique Shadow Fading Fast Fading Received Signal Strength Bit error rate Packet error rate Fade Margin Increase transmit power or decrease cell size Antenna Diversity Error control coding Interleaving Frequency enc hopping Multipath Delay Spread Inter-symbol Interference Adaptive Equalization DS-Spread Spectrum OFDM Directional Antennas Telcom 2700 28 Error Control BER in wireless networks Several orders of magnitude worse than wireline networks (eg, 10-2 vs 10-10 in optic fibers) Channel errors are random and bursty, usually coinciding with deep fast fades Much higher BER within bursts Protection against bit errors Necessary for data Speech can tolerate much higher bit errors (< 10-2 depending on encoding/compression algorithm) Error Control Coding used to overcome BER Telcom 2700 29

Error control coding Coding is a form of diversity Transmit redundant bits from which you can detect/recover from errors The redundant bits have a pattern that enables this recovery Approaches to error control Error Detection + ARQ Error Correction (FEC) k bit data block Simple Block code Block Encoder n k parity check bits n bit codeword k data bits Telcom 2700 30 Error control Error control coding: systematically add redundant bits for error detection or correction Error detection codes: Detect whether received word is a valid codeword but not enough redundancy to correct bits Retransmit data after error (automatic repeat request ARQ) Error correction codes (forward error correction: FEC) Detect invalid codewords and correct into valid codeword Correction requires more bits than error detection FEC is good for one-way channels, recordings (CD-ROMs), real-time communications, deep space,... Generally more bits are required to protect against larger number of bit errors Telcom 2700 31

Single Parity Example: single parity bit even parity code Valid codewords should always have even number of 1 s Add a parity bit=1 if number of 1 s in data is odd add parity bit=0 if number of 1 s in data is even If any bit is errored, the received codeword will have odd number of 1 s Single parity can detect any single bit error (but not correct) Actually, any odd number of bit errors can be detected Telcom 2700 32 Single Parity (cont) Example 3 bits parity bit --> 2 3 valid codewords transmission received codewords valid invalid 0000 0001 0011 0010 0101 0100 0110 0111 1001 1000 1010 1011 1100 1101 1111 1110 Single bit error will change valid word into an invalid word (detectable); double bit error will change valid word into another valid word (undetectable) Telcom 2700 33

(n,k) block codes Block Codes k = number of data bits in block (data word length) n-k = number of parity check bits added n = length of codeword or code block (n-k)/n = overhead or redundancy (lower is more efficient) k/n = coding rate (higher is more efficient) Telcom 2700 34 Block Codes (cont) data k bits n bits data data parity check transmission 2 n possible codewords, only 2 k are valid n-bit codeword --> 2 k valid codewords valid? error no bit errors Telcom 2700 35

Block Code Principles Hamming distance : for 2 n-bit binary sequences, the number of different bits e.g., v 1 =011011; v 2 =110001; d(v1, v 2 )=3 The minimum distance (d min ) of an (n,k) block code is the smallest Hamming distance between any pair of codewords in a code. Number of error bits can be detected: d min-1 Number of error bits can be corrected t: t d min 2 1 Telcom 2700 36 (7,4) Hamming code Message word Code word Weight 0000 000 0000 0 0001 101 0001 3 0010 111 0010 4 0011 010 0011 3 0100 011 0100 3 0101 110 0101 4 0110 100 0110 3 0111 001 0111 4 1000 110 1000 3 1001 011 1001 4 1010 001 1010 3 1011 100 1011 4 1100 101 1100 4 1101 000 1101 3 1110 010 1110 4 1111 111 1111 7 2 7 possible 7-bit words (128 possible) of which we use only 16 All codewords are distance 3 apart => Can detect 2 errors, correct 1 error Telcom 2700 37

Forward Error Correction Process FEC Operation Transmitter Forward error correction (FEC) encoder maps each k-bit block into an n-bit block codeword Codeword is transmitted; Receiver Incoming signal is demodulated Block passed through an FEC decoder Decoder detects and correct errors Receiver can correct errors by mapping invalid codeword to nearest valid codeword Telcom 2700 38 FEC (cont) Decoding sphere valid codeword C1 distance e distance d 2e+1 valid codeword C2 distance e Telcom 2700 39

Convolutional Codes Block codes treat data as separate blocks (memoryless encoding); Convolutional codes map a continuous data string into a continuous encoded string (memory) Error checking and correcting carried out continuously (n, k, K) code Input processes k bits at a time Output produces n bits for every k input bits K = constraint factor k and n generally very small n-bit output of (n, k, K) code depends on: Current block of k input bits Previous K-1 blocks of k input bits Telcom 2700 40 Convolutional Codes (cont) Successive k-tuples are mapped into n-tuples n-tuples should be designed to have distance properties for error detection/correction Example: K=3 stages, k=1, n=2 bits output input data bit bit bit + + first bit second bit Telcom 2700 41

Convolutional Encoder Can represent coder by state transition diagram with 2 (K-1) states Telcom 2700 42 What does coding get you? Consider a wireless link probability of a bit error = q probability of correct reception = p In a block of k bits with no error correction P(word correctly received) = p k P(word error) = 1 p k With error correction of t bits in block of n bits t n n i i P( word correct ) ( p) q P( word i0 i error ) 1 P( word correct ) Telcom 2700 43

What does coding get you? Example consider (7,4) Hamming Code when BER = q =.01, p =.99 In a block of 4 bits with no error correction P(word correctly received) = p k =.9606 P(word error) = 1 p k = 0.04 With error correction of 1 bits in block of 7 bits P( word P( word t n ni i 7 7 correct ) ( p) q p ( p) i0 i 6 error ) 1 P( word correct ) 0.002 7 q 1 0.998 Get an order of magnitude improvement in word error rate Telcom 2700 44 Interleaving Problem: Errors in wireless channels occur in bursts due to fast fades Error correction codes designed to combat random errors in the code words Interleaving Idea: If the errors can be spread over many codewords they can be corrected- achieved my shuffling codewords makes the channel memoryless and enables coding schemes to perform in fading channels. the penalty is the delay in receiving information- bits have to be buffered for interleaving Interleaving is performed after coding at receiver deinterleave before decoding Telcom 2700 45

Block interleaving After codewords are created, the bits in the codewords are interleaved and transmitted This ensures that a burst of errors will be dispersed over several codewords and not within the same codeword Needs buffering at the receiver to create the original data The interleaving depth depends on the nature of the channel, the application under consideration, etc. codeword Direction of transmission Bits in error dispersed over several codewords Telcom 2700 46 Interleaving Example Usually transmit data in order it arrives bit bit position : : 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 a 0 a 1 a 2 a 3 a 4 a 5 a 6 b 0 b 1 b 2 b 3 b 4 b 5 b 6 c 0 c 1 c 2 c 3 Suppose bits 6 to 11 are in error because of a fade The codewords a and b are lost. Suppose we interleaving at depth 7 by buffering up 7 words then output them in order to bit positions of the words bit position bit : : 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 a 0 b 0 c 0 d 0 e 0 f 0 g 0 a 1 b 1 c 1 d 1 e 1 f 1 g 1 a 2 b 2 c 2 d 2 Now can correct a fade that results in bits 6-11 being lost Telcom 2700 47

Frequency Hopping Traditionally: transmitter/receiver pair communicate on fixed frequency channel. Frequency Hopping Idea: Since noise, fading and interference change somewhat with frequency band used move from band to band Time spent on a single frequency is termed the dwell time Originally for military communications Spend a short amount of time on different frequency bands to prevent interception or jamming - developed during WWII by actress Hedy Lammar and classical composer George Antheil patent given to government Telcom 2700 48 Frequency Hopping concept f f 8 f 7 f 6 f 5 f 4 f 3 f 2 f 1 time Timeslot for transmission Unacceptable errors Telcom 2700 49

Frequency Hopping Two types: Slow Hopping Dwell time long enough to transmit several bits in a row (timeslot) Fast Hopping Dwell time on the order of a bit or fraction of a bit (primarily for military systems) Transmitter and receiver must know hopping pattern/ algorithm before communications. Cyclic pattern best for low number of frequencies and combating Fast Fading : Example with four frequencies: f4, f2, f1, f3, f4, f2, f1, f3,. Random pattern best for large number of frequencies, combating co-channel interference, and interference averaging Example with six frequencies: f1, f3, f2, f1, f6, f5, f4, f2, f6, Use random number generator with same seed and both ends Telcom 2700 50 Frequency Hopping Slow frequency hopping used in cellular (GSM) Fast in WLANs Provides interference averaging and frequency diversity By hopping mobile less like to suffer consecutive deep fades Telcom 2700 51

Multipath propagation Signal can take many different paths between sender and receiver due to reflection, scattering, diffraction signal at sender signal at receiver Time dispersion: signal is dispersed over time interference with neighbor symbols, Inter Symbol Interference (ISI) The signal reaches a receiver directly and phase shifted distorted signal depending on the phases of the different parts Telcom 2700 52 Equalization Equalizer filter that performs the inverse of the channel to compensate for the distortion created by multipath delay (combats ISI) In wireless networks equalizers must be adaptive channel is usually unknown and time varying equalizers track the time variation and adapt Two step approach to equalization 1. Training: a known fixed-length sequence is transmitted for the receiver s equalizer to train on that is to set parameters in the equalizer 2. Tracking: the equalizer tracks the channel changes with the help of the training sequence, and uses a channel estimate to compensate for distortions in the unknown sequence. Telcom 2700 53

Adaptive Equalization Training step, the channel response, h(t) is estimated Tracking step, the input signal, s(t), is estimated Equalizers are used in NA-TDMA, GSM and HIPERLAN There are several different types of equalizers (3 popular ones) LTE: Linear transversal filter DFE: Decision feedback equalizer MLSE: Maximum likelihood sequence estimators Disadvantages of equalizers are complexity, bandwidth consumption and power consumption Telcom 2700 54 Adaptive Equalizers Typical equalizer implemented as a tap delay line filter with variable tap gains Telcom 2700 55

Spread Spectrum Radio Aspects Military Spread spectrum techniques adapted for cellular systems 1. Frequency Hopping: vary frequency transmit on 2. Direct Sequence Narrowband signal is spread over very large bandwidth signal using a spreading signal Spreading signal is a special code sequence with rate much greater than data rate of message The receiver uses correlation to recover the original data Multipath fading is reduced by direct sequence signal spreading and better noise immunity DS also allows lower power operation Telcom 2700 56 Direct Sequence Spread Spectrum Each bit in original signal is represented by multiple bits in the transmitted signal Spreading code spreads signal across a wider frequency band Spread is in direct proportion to number of chip bits W used Processing gain G = W/R; W = chips per sec, R = information bit rate per sec Processing gain is a measure of the improvement in SNR gained by using the additional bandwidth from spreading (18-23 db in cellular l systems) One Spreading technique combines digital information stream with the spreading code bit stream using exclusive- OR Telcom 2700 57

DSSS Modulation The original data stream is chipped up into a pattern of pulses of smaller duration Good correlation properties Good crosscorrelation properties with other patterns Each pattern is called a spread spectrum code Instead of transmit 1 data bit transmit 11 chip bit pattern - adding redundancy Data Bit 1 2 3 4 5 6 7 8 9 10 11 Data In Spreading Code In Spread Bits chip Periodic Spreading Code Telcom 2700 58 DSSS (Direct Sequence Spread Spectrum) II user data X spread spectrum signal modulator transmit signal chipping sequence radio carrier transmitter correlator received signal demodulator lowpass filtered signal X products integrator sampled sums data decision radio carrier chipping sequence receiver Telcom 2700 59

DSSS Spectrum amplitude 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 original spectrum spread spectrum 0-10 -8-6 -4-2 0 2 4 6 8 10 normalized frequency spectrum Example: IEEE 802.11 Wi-Fi Wireless LAN standard Uses DSSS with 11 bit chipping code To transmit a 0, you send [1 1 1-1 -1-1 1-1 -1 1-1] To transmit a 1 you send [-1-1 -1 1 1 1-1 1 1-1 1] Processing gain The duration of a chip is usually represented by T c The duration of the bit is T The ratio T/T c = R is called the processing gain of the DSSS system For 802.11 R = 11 Telcom 2700 60 Example in a two-path channel Random data sequence of ten data bits Spreading by 11 chips using 802.11b chipping code Two path channel with interpath delay of 17 chips > bit duration Multipath amplitudes Main path: 1 Second path: 1.1 Reality: Many multipath components Different path amplitudes Noise 2 1.5 1 0.5 0-0.5-1 -1.5-2 0 0 0 0 1 1 0 0 1 0 10 20 30 40 50 60 70 80 90 100 110 Telcom 2700 61

Output without spreading Errors introduced by the channel 15 25 10 0 0 0 0 1 1 0 0 1 0 20 15 0 0 0 0 1 1 0 0 1 0 5 10 0 5-5 0-5 -10-10 -15 0 20 40 60 80 100 120-15 0 20 40 60 80 100 120 Without Multipath With Multipath Telcom 2700 62 Output with spreading Output of DSS demodulator Errors introduced by the channel are removed 15 15 10 0 0 0 0 1 1 0 0 1 0 0 0 0 0 1 1 0 0 1 0 10 5 5 0 0-5 -5-10 -10-15 0 20 40 60 80 100 120 Without Multipath -15 0 20 40 60 80 100 120 With Multipath Telcom 2700 63

Performance Degradation and Diversity Issue Performance Affected Diversity Technique Shadow Received Signal Strength Fade Margin Increase Fading transmit power or decrease cell size Fast Fading (Time Variation) Multipath Delay Spread (Time Dispersion) Bit error rate Packet error rate Inter-symbol Interference Antenna Diversity Error control coding Interleaving Frequency hopping Adaptive Equalization DS-Spread Spectrum OFDM Directional Antennas Telcom 2700 64 Typical Wireless Communication System Source Source Encoder Channel Encoder Modulator Channel Destination Source Decoder Channel Decoder Demod -ulator Telcom 2700 65

Source Coding Source Coding seeks to efficiently encode the information for transmission Mobile environment much different from wired issues Efficient use of spectrum Compress to lower bit rate per user => more users Quality Want near tollgrade in a very difficult transmission environment (high BER, bursty errors) Hardware complexity Size, speed and power consumption Consider Digital Speech as example Convert speech to digital form and transmit digitally Telcom 2700 66 Digital Speech Speech Coder: device that converts speech to digital Types of speech coders Waveform coders Convert any analog signal to digital form Vocoders (Parametric coders) Try to exploit special properties of speech signal to reduce bit rate Build model of speech transmit parameters of model Hybrid Coders Combine features of waveform and vocoders Telcom 2700 67

Waveform Coders (e.g.,pcm) Waveform Coders Convert any analog signal to digital - basically A/D converter Analog signal sampled > twice highest frequency- then quantized into ` n bit samples Uniform quantization Example Pulse Code Modulation band limit speech < 4000 Hz pass speech through law compander sample 8000 Hz, 8 bit samples 64 Kbps DS0 rate Characteristics Quality High Complexity Low Bit rate High Delay - Low Robustness - High Telcom 2700 68 Characteristics of Speech Bandwidth Most of energy between 20 Hz to about 7KHz, Human ear sensitive to energy between 50 Hz and 4KHz Time Signal High correlation Short term stationary Classified into four categories Voiced : created by air passed through vocal cords (e.g., ah, v) Unvoiced : created by air through mouth and lips (e.g., s, f ) Mixed or transitional Silence Telcom 2700 69

Characteristics of Speech Typical Voiced speech Typical Unvoiced speech Telcom 2700 70 Vocoders Vocoders (Parametric Coders) Models the vocalization of speech Speech sampled and broken into frames (~25 msec) Instead of transmitting digitized speech build model of speech transmit parameters of model and synthesize approximation of speech Linear Predictive Coders (LPC) Models Vocal tract as a filter Filter excitation periodic pulse (voiced speech) noise (unvoiced) Transmitted parameters: gain, voiced/unvoiced decision, pitch (if voiced), LPC parameters Telcom 2700 71

Vocoders Example Tenth Order Linear Predictive Coder Samples Voice at 8000 Hz buffer 240 samples => 30 msec Filter Model (M=10 is order, G is gain, z -1 unit delay, b k are filter coefficients) H ( z ) 1 M G k 1 b k z k G = 5 bits, b k = 8 bits each, voiced/unvoiced decision = 1 bit, pitch = 6 bits => 92 bits/30 msec = 3067 bps Telcom 2700 72 Vocoders LPC coders can achieve low bit rates 1.2 4.8 Kbps Characteristics of LPC Quality Low Complexity Moderate Bit Rate Low Delay Moderate Robustness Low Quality of pure LPC vocoder to low for cellular telephony - try to improve quality by using hybrid coders Hybrid Coders Combine Vocoder and Waveform Coder concept Residual LPC (RELP) Codebook excited LPC (CELP) Telcom 2700 73

RELP Vocoder Residual Excited LPC improve quality of LPC by transmitting error (residue) along with LPC parameters s(n) Buffer/ Window residue ST-LP Analysis v/u decision gain, pitch LP parameters LPC Synthesis ENCODER Encoded output Block diagram of a RELP encoder Telcom 2700 74 GSM Speech Coding Analog speech Low-pass filter A/D 104 kbps 13 kbps RPE-LTP Channel speech encoder encoder 8000 samples/s, 13 bits/sample Telcom 2700 75

GSM Speech Coding (cont) Regular pulse excited - long term prediction (RPE-LRP) speech encoder (RELP speech coder 160 samples/ 20 ms from A/D (= 2080 bits) RPE-LTP speech encoder 36 LPC bits/20 ms 9 LTP bits/5 ms 47 RPE bits/5 ms 260 bits/20 ms to channel encoder LPC: linear prediction coding filter LTP: long term prediction pitch + input RPE: Residual Prediction Error: Telcom 2700 76 Hybrid Vocoders Codebook Excited LPC Problem with simple LPC is U/V decision and pitch estimation doesn t model transitional speech well, and not always accurate Codebook approach pass speech through an analyzer to find closest match to a set of possible excitations (codebook) Transmit codebook pointer + LPC parameters 3G, ITU G.729 standard Telcom 2700 77

CELP Speech Coders General CELP architecture Telcom 2700 78 Evaluating Speech Coders Qualitative Comparison based on subjective procedures in ITU-T Rec. P. 830 Major Procedures Absolute Category Rating Subjects listen to samples and rank them on an absolute scale - result is a mean opinion score (MOS) Comparison Category Rating Subjects listen to coded samples and original uncoded sample (PCM or analog), the two are compared on a relative scale result is a comparison mean opinion score (CMOS) Mean Opinion Score (MOS) ------------------------------------- Excellent 5 Good 4 Fair 3 Poor 2 Bad 1 Comparison MOS (CMOS) ------------------------------------- Much Better 3 Better 2 Slightly Better 1 About the Same 0 Slightly Worse -1 Worse -2 Much Worse -3 Telcom 2700 79

Evaluating Speech Coders MOS for clear channel environment no errors Result vary a little with language and speaker gender Standard Speech coder Bit rate MOS PCM Waveform 64 Kbps 4.3 CT2 Waveform 32 Kbps 4.1 DECT Waveform 32 Kbps 4.1 NA-TDMA Hybrid CELPC 8Kbps 3.0 GSM Hybrid RELPC 13 kbps 3.54 QCELP Hybrid CELPC 14.4 Kbps 3.4 4.0 QCELP Hybrid CELPC 9.6 Kbps 3.4 LPC Vocoder 2.4 Kbps 2.5 ITU G.729 Hybrid CELP 8.Kbps 3.9 Telcom 2700 80 Multiple Access and Mode Mode Simplex one way communication (e.g., broadcast AM) Duplex two way communication TDD time division duplex users take turns on the channel FDD frequency division duplex users get two channels one for each direction of communication For example one channel for uplink (mobile to base station) another channel for downlink (base station to mobile) Multiple Access determines how users in a cell share the frequency spectrum assigned to the cell: FDMA, TDMA, CDMA Wireless systems often use a combination of schemes; GSM FDD/FDMA/TDMA Telcom 2700 81

Multiple Access Techniques FDMA (frequency division multiple access) separate spectrum into non-overlapping frequency bands assign a certain frequency to a transmission channel between a sender and a receiver different users share use of the medium by transmitting on non-overlapping frequency bands at the same time TDMA (time division multiple access): assign a fixed frequency to a transmission channel between a sender and a receiver for a certain amount of time (users share a frequency channel in time slices) CDMA (code division multiple access): assign a user a unique code for transmission between sender and receiver, users transmit on the same frequency at the same time Telcom 2700 82 Frequency division multiple access frequency Telcom 2700 time

Time Division Multiple Access frame slot frequency time Telcom 2700 84 code Code Division Multiple Access time frequency Telcom 2700 85

FDMA FDMA is simplest and oldest method Bandwidth F is divided into T non-overlapping frequency channels Guard bands minimize interference between channels Each station is assigned a different frequency Can be inefficient if more than T stations want to transmit or traffic is bursty (resulting in unused bandwidth and delays) Receiver requires high quality filters for adjacent channel rejection Used in First Generation Cellular (NMT) f1 f2 Telcom 2700 86 FDD/FDMA - general scheme, example AMPS (B block) f 893.97MHz 799 880.65 MHz 849.97 MHz 355 799 20 MHz 30 khz 835.65 MHz 355 t f(c) = 825,000 + 30 x (channel number) KHz <- uplink f(c) = f uplink + 45,000 KHz <- downlink In general all systems use some form of FDMA Telcom 2700 87

TDMA Users share same frequency band in nonoverlapping time intervals, eg, by round robin Receiver filters are just windows instead of bandpass filters (as in FDMA) Guard time can be as small as the synchronization of the network permits All users must be synchronized with base station to within a fraction of guard time Guard time of 30-50 microsec common in TDMA Used in GSM, NA-TDMA, (PDC) Pacific Digital Cellular Telcom 2700 88 TDD/TDMA - general scheme, example 417 µs 1 2 3 11 12 1 2 3 11 12 downlink uplink t CT2 cordless phone standard Telcom 2700 89

Air Interface 25 MHz of bandwidth is divided into 124 frequency bands of 200 khz each and two 100 khz pieces on either side Carrier frequencies are given by: Fu (n) = 890.2 + 0.2(n-1) MHz n=1,2,3,,124 Fd (n) = 935.2 + 0.2(n-1) MHz n=1,2,3,,124 Example: On the uplink, Channel 1 = 890.1-890.3 MHz On the downlink, Channel 1 = 935.1-935.3 MHz Usually, Channels 1 and 124 will not be used if possible Telcom 2700 90 CDMA Code Division Multiple Access Narrowband message signal is multiplied by very large bandwidth spreading signal using direct sequence spread spectrum All users can use same carrier frequency and may transmit simultaneously Each user has own unique access spreading codeword which is approximately orthogonal to other users codewords Receiver performs time correlation operation to detect only specific codeword, other users codewords appear as noise due to decorrelation Cocktail party example Telcom 2700 91

CDMA and direct sequence spread spectrum The original data stream is V Data Bit chipped up into a pattern of 1 1 2 3 4 5 6 7 8 9 10 11 0.9 pulses of 0.8 smaller duration t 0.7 Each pattern is 0.6 called a spread 0.5 spectrum code Spread Bit 0.4 The primary V 0.3 spread 0.2 spectrum code 0.1 used in CDMA t is the Walsh Code chip amplitude original spectrum spread spectrum 0-10 -8-6 -4-2 0 2 4 6 8 10 normalized frequency Telcom 2700 92 Simple example illustrating CDMA Traditional To send a 0, send +1 V for T seconds To send a 1, send -1 V for T seconds Use separate time slots or frequency bands to separate signals Simple CDMA To send a 0, Bob sends +1 VforT seconds; Alice sends +1 V for T/2 seconds and -1 V for T/2 seconds To send a 1, Bob sends -1 V for T seconds; Alice sends -1 V for T/2 seconds and +1 V for T/2 Data 1 V 0 1 1 0 1 T T time T T chip Code [1, 1] [-1, -1] [1, -1] [-1, 1] Telcom 2700 93

Simple CDMA Transmitter User 1 data in Spread User 2 data in Spread V 1 0011 0 1 = [1, 1, 1, 1, -1, -1, -1, -1] T 2T 3T 4T t 2 V V 1 0 1 0 = [-1, 1, 1, -1, -1, 1, 1, -1] 1 Transmitted signal T 2T 3T 4T Telcom 2700 94 t T 2T 3T 4T t Simple CDMA Receiver Despread correlate with [1, 1] User 1 data out Despread User 2 data out Received signal 2 1 V V T 2T 3T 4T correlate with [1, -1] Alice s Code t = T 2T 3T 4T -T has a negative sign Alice sent a 1 as the first bit Telcom 2700 95 t = -2 x T/2 = -T

Simple CDMA continued Proceeding in this fashion for each bit, the information transmitted by Alice can be recovered To recover the information transmitted by Bob, the received signal is correlated bit-by-bit with Bob s code [1,1] Such codes are orthogonal Multiply the codes element-wise [1,1] x [1,-1] = [1,-1] Add the elements of the resulting product 1 + (-1) = 0 => the codes are orthogonal In CDMA (IS-95) each orthogonal code is called a Walsh Code and has 64 chips in it Note that the transmissions MUST be synchronized to recover data Otherwise there will be interference and errors Telcom 2700 96 CDMA Properties: Near-Far Problem A CDMA receiver cannot successfully despread the desired signal in a high multiple-access-interference environment Unless a transmitter close to the receiver transmits at power lower than a transmitter farther away, the far transmitter cannot be heard Power control must be used to mitigate the near-far problem Mobile transmit so that power levels are equal at base station Base station Telcom 2700 97

Summary Fundamental Concepts for Wireless Communications Modulation Techniques Digital Modulation Diversity i Techniques Error Control Coding Interleaving Adaptive Equalizers Frequency Hopping Spread Spectrum Direct Sequence Spread Spectrum Mode and Multiple l Access TDD/FDD FDMA TDMA CDMA Telcom 2700