1. Overview
1.1 Definition and Main Characteristics
- Definition: A mobile communication link involves at least one communicating party in motion. It covers fixed-to-mobile, mobile-to-mobile, and person-to-person mobile scenarios.
- Key traits: Uses radio waves, operates in heavy-interference environments (Doppler, shadowing, multipath), works with scarce spectrum, complicates network management because of mobility, and demands high-performance user equipment.
1.2 Development Trends
- Mobile subscriptions outpace fixed lines and drive the ICT industry.
- Mobile data traffic keeps rising; mobile Internet access becomes essential.
- Terminals evolve toward intelligent, broadband, standardized devices that become primary Internet access points.
- Networks trend toward convergence, intelligence, and global coverage.
2. Wireless Channel
2.1 Propagation Characteristics
- Propagation modes: Received signals combine direct, reflected, diffracted, and scattered components.
- Impairments: Path loss grows with distance and frequency; terrain/obstacles introduce shadow fading; multipath superposition causes random amplitude/phase/delay variations (multipath fading); motion induces Doppler shifts that broaden spectra; other transmitters and environmental noise add interference.
- Mobile channel: A time-varying channel whose fading profile results from the combined effects of path loss, shadowing, and small-scale fading.
2.2 Fading
- Definition: Rapid fluctuations of received field strength caused by complex propagation paths.
- Classification: Frequency-selective vs. flat fading; fast vs. slow fading.
- Effects: Lower received level, waveform distortion, delay spread that breaks synchronization, and tracking challenges under fast fading.
-
Large-scale (slow) fading: Caused by path loss and shadowing. Path loss increases with distance and carrier frequency. Example outdoor models: Okumura (1500–1920 MHz macro cells) and Hata (150–1500 MHz, radius > 1 km). Indoor: Keenan–Motley. Shadowing produces log-normal envelopes:
\[P(r)=\frac{1}{\sqrt{2\pi\mu_s}}\exp\left[-\frac{(r-m)^2}{2\mu_s^2}\right]\]Mitigation: deploy additional base stations.
- Small-scale (fast) fading: Driven by multipath. Envelopes typically follow Rayleigh or Rician distributions. Time dispersion is described by mean excess delay and RMS delay spread; coherence bandwidth satisfies \(\Delta f=1/(6\pi\sigma_\tau)\) for correlation 0.9 or \(\Delta f=1/(2\pi\sigma_\tau)\) for 0.5. Larger delay spreads narrow \(\Delta f\) and exacerbate intersymbol interference (ISI). Coherence time relates to Doppler spread by \(T_c=\sqrt{9/(16\pi f_m^2)}\). Mitigation: diversity reception.
3. Voice Coding, Channel Coding, and Modulation
3.1 Voice Coding
- Compression rationale:
- External: Speech has redundancy—non-uniform amplitude distribution and strong short/long-term correlation.
- Internal: Human auditory masking tolerates quantization noise and phase distortion.
- Categories:
- Waveform coders (PCM, ADPCM) keep high fidelity at higher bitrates.
- Parametric coders (vocoder/LPC-10) encode model parameters at low bitrates with poorer quality.
- Hybrid coders (CELP) blend both ideas for better quality at low rates.
- Requirements: Low rate (<16 kbps pure coding), high quality, low delay (tens of ms), robustness to channel errors, moderate complexity.
- Key techniques: Short-time analysis (frames of 10–30 ms, often via HMMs), pitch estimation (autocorrelation/average magnitude difference), linear prediction to remove short-term correlation, and vector quantization for residuals.
- CELP: Uses analysis-by-synthesis with both fixed and adaptive codebooks. Frames (20–30 ms) split into subframes; the best excitation vector per subframe is chosen. Perceptual weighting shapes noise using auditory masking. Redundancy removal spans intra-frame (LP) and inter-frame (adaptive codebook) correlations; remaining residuals are approximated by the fixed codebook.
- Quality evaluation: Subjective MOS scores (1–5) or objective metrics such as SNR and PESQ.
3.2 Channel Coding
- Definition: Adds parity symbols to satisfy constraints so the codeword can withstand channel impairments.
- Purpose: Improve transmission reliability.
- Classification: Block vs. convolutional; linear vs. nonlinear; error-detecting vs. error-correcting.
- Block codes: Expand k-bit messages to n-bit codewords by inserting n−k parity bits.
- Convolutional codes: Simple encoders with strong performance but complex decoders; excel at random errors but not bursts.
- Interleaving: Rearranges coded bits to turn burst errors into random ones; not a code itself but prepares data for random-error-correcting codes.
- Turbo codes: Parallel concatenation of two convolutional encoders separated by an interleaver; outputs systematic data plus two parity streams for an effective 1/3 rate. Decoding iteratively exchanges soft information to approach Shannon limits.
- LDPC codes: Sparse linear block codes that offer near-capacity performance with low-complexity, parallelizable decoding and built-in error detection.
- Polar codes: Capacity-achieving block codes based on channel polarization.
3.3 Modulation
-
Signal-space analysis: Digital symbols map to waveform vectors via orthonormal basis functions such as
\[\phi_1(t)=\sqrt{\frac{2}{T}}\cos(2\pi f_ct),\quad \phi_2(t)=\sqrt{\frac{2}{T}}\sin(2\pi f_ct)\]Each signal \(s_i(t)\) is represented by coefficient vector \(s_i\), and constellation distances follow
\[\|s_i-s_k\|=\sqrt{\sum_{j=1}^N(s_{ij}-s_{kj})^2}=\sqrt{\int_0^T (s_i(t)-s_k(t))^2 dt}.\] - Modulation families: Non-constant envelopes (ASK, QAM, APSK) and constant envelopes (FSK, PSK, CPM). PSK/QAM constellations illustrate symbol placement.
- MSK: Minimum shift keying with constant envelope, narrow bandwidth, coherent demod capability but strict adjacent-channel interference limits.
- GMSK: Gaussian-filtered MSK that shapes data before modulation.
- Multicarrier modulation: Splits data into many low-rate substreams, each modulating an orthogonal subcarrier so every subchannel is narrower than the coherence bandwidth.
- OFDM: Applies parallel orthogonal subcarriers plus cyclic prefixes to convert a frequency-selective channel into flat subchannels. Advantages: high spectral efficiency, FFT-based implementation, supports adaptive modulation, asymmetric services. Drawbacks: sensitivity to frequency offsets/phase noise (ICI) and high PAPR.
4. Performance Enhancement Techniques
4.1 Spread Spectrum
- Features: Uses bandwidth far exceeding the original signal via high-rate pseudo-random codes. Provides resilience to intentional and narrowband interference and mitigates multipath.
- Shannon rationale: For \(S/N \ll 1\), \(C/B \approx 1.44 S/N\), so increasing bandwidth allows reliable communication at low SNR. After despreading, desired signals collapse back into a narrow band while interference remains dispersed.
- DSSS: Multiply the data stream by a pseudo-random code, modulate, transmit, then despread using a synchronized local code before demodulation.
- FHSS: Drive the RF oscillator with a PN sequence so the carrier hops rapidly; effectively a sequence of narrowband transmissions over a wide spectrum.
- Codes: Ideal codes provide many sequences, sharp autocorrelation, zero cross-correlation, and high complexity. Walsh codes are orthogonal yet unsuitable as spread-spectrum codes because different codewords occupy unequal bandwidths. PN codes (m-sequences, Gold sequences) approximate noise and work well.
4.2 Diversity
- Concept: Send the same data over independent fading paths so the probability of simultaneous deep fades is low, improving SNR by 10–20 dB after combining.
- Classes: Micro-diversity combats multipath fading; macro-diversity handles shadowing.
- Implementations:
- Spatial diversity via antenna arrays (spacing ≈ half wavelength).
- Polarization diversity (vertical/horizontal) with up to two branches and a 3 dB split loss.
- Angle diversity using directional beams.
- Frequency diversity with carriers separated beyond coherence bandwidth (often using RAKE).
- Time diversity via repeated transmissions or coding/interleaving beyond coherence time.
- Combining: Selection combining (pick best branch), equal-gain combining, and maximal ratio combining (weight by SNR), with performance ordered MRC > EGC > SC.
- RAKE receivers: Use multiple correlators aligned to different multipath delays, then align and combine. Variants include A-RAKE (all paths), S-RAKE (strongest L paths), and P-RAKE (earliest L paths).
4.3 MIMO
- Definition: Multiple antennas at both transmitter and receiver create several parallel spatial channels.
- Modes: Spatial multiplexing (split data into low-rate streams and separate them via detection), spatial diversity (orthogonal encoding across antennas), and beamforming (focus energy using channel knowledge to boost SNR and reduce interference).
- Core techniques: Space-time processing, including space-time trellis codes (STTC) and space-time block codes (STBC).
4.4 Equalization
- Time-domain equalizers: Manipulate impulse responses directly, using linear or nonlinear structures, to meet ISI-free conditions.
- Frequency-domain equalizers: Shape amplitude and group delay of the overall response (channel × equalizer).
- Adaptive equalization: Continuously tunes coefficients via training and tracking modes to follow channel variations.
5. Networking Technologies
5.1 Cellular Architecture
- Cell types: Belt-shaped service areas (two- or three-frequency) and planar honeycomb layouts. Advantages include higher frequency reuse, flexible deployment, and complex topology.
- Clusters: \(N_R\) cells sharing the full band form a cluster (size \(N_R\)). Cells inside a cluster use distinct frequencies; reuse occurs only between different clusters. Adjacent clusters must maintain equal co-channel spacing.
- Reuse factor: \(Q=D/R=\sqrt{3N_R}\), where \(D\) is co-channel distance and \(R\) cell radius. Larger \(Q\) reduces interference but lowers capacity.
-
Carrier-to-interference ratio:
\[\frac{S}{I}=\frac{R^{-n}}{\sum_{i=1}^{i_0}D_i^{-n}}\]Assuming equal-distance first-tier interferers, \(S/I = (\sqrt{3N})^n/i_0\).
- Excitation: Center-fed or vertex-fed cells.
- Cell splitting: Decrease cell radius or increase per-cell channels to boost capacity in dense areas.
5.2 Multiple Access
- Definition: Assign distinguishing signatures so multiple users share a common medium, raising spectral efficiency.
- Types: FDMA, TDMA, CDMA, SDMA.
- FDMA: Divides total bandwidth into disjoint subbands (channels) with guard bands; suitable for narrowband systems, requires band-pass filters, and complicates handover. Interference sources include adjacent-channel, co-channel, and intermodulation.
- TDMA: Partitions channels by time slots (see frame diagram). Offers high burst rates, removes duplexer needs, simplifies handover, but requires adaptive equalization at high rates.
- CDMA: Assigns quasi-orthogonal PN codes per user. Handles multiple users on the same spectrum, offers soft capacity, mitigates multipath, supports smooth soft handover and macro diversity, but must address multiple-access interference and near-far effects.
- SDMA: Uses adaptive antenna arrays to form beams toward different users, boosting capacity, coverage, compatibility, and localization while lowering interference and power, but must be combined with other MA schemes.
- System capacity: \(m = \frac{B_t}{B_c N} = \frac{B_t}{B_c\sqrt{\frac{2}{3}\left(\frac{C}{I}\right)_{min}}}\).
6. GSM
6.1 System Overview
- Services:
- Bearer services (restricted voice, async/sync duplex data, packet options).
- Telecommunication services (voice, emergency, SMS, fax).
- Supplementary services such as caller ID, conference calling, charging alerts.
- Architecture:
- Mobile Station (MS): Mobile terminal plus SIM.
- Base Station Subsystem (BSS): BTS, BSC, and PCU handling radio coverage and resources.
- Network and Switching Subsystem (NSS): MSC, VLR, HLR, EIR, AUC for routing, management, security, and mobility.
- Operation Support Subsystem (OSS): OMC and peripherals for O&M tasks.
6.2 Channels
- Time-slot/frame hierarchy: Slots (0.577 ms) form TDMA frames (8 slots, 4.615 ms), which compose 26-frame traffic multi-frames or 51-frame control multi-frames; superframes contain 51 traffic or 26 control multi-frames (1326 frames, 6.12 s); hyperframes contain 2048 superframes (3 h 28 min 53.760 s).
- Logical channels: Traffic channels (full/half-rate voice or data) plus control channels (BCH, CCCH, DCCH).
- Mapping: Each RF carrier offers eight physical channels but more logical channels, so multiplexing is required. Example: carrier \(C_0\) uses TS0–TS1 for control, TS2–TS7 for traffic; other carriers devote all slots to traffic.
- Burst types: Normal, frequency-correction, synchronization, access, and idle bursts.
6.3 Radio Transmission Technologies
- Speech codec: RPE-LTP (Regular Pulse Excited with Long-Term Prediction) at 8 kHz sampling, 20 ms frames, 260 bits per frame (13 kbps).
- Channel coding / interleaving: See block diagram; GSM applies two-stage interleaving (intra-frame then block interleaving).
- Modulation: GMSK.
- Other techniques: Fading mitigation (diversity, adaptive filtering, frequency hopping) and power-saving features (adaptive power control, discontinuous transmission/reception).
6.4 Control and Management
- Covers registration, roaming, and handover procedures.
6.5 GPRS
- Packet-switched enhancement that avoids permanent connections between the mobile station and external networks, yielding higher data efficiency.
6.6 EDGE
- Enhanced Data rate for GSM Evolution introduces 8PSK modulation, new MCS coding, link adaptation, incremental redundancy ARQ, updated RLC/MAC, and improved link-quality monitoring.
7. CDMA IS-95
7.1 System Overview
- Uses three code families: short PN codes (cell identification), Walsh codes (forward-channel separation), and long PN codes (reverse-link user separation).
7.2 Forward Link
- Downlink supports up to 64 simultaneous channels on one RF carrier, distinguished by orthogonal Walsh sequences.
- Pilot channel: Unmodulated DSSS signal for timing, coherent demodulation, and handover decisions.
- Sync channel: Broadcasts synchronization data.
- Paging channel: Broadcasts system parameters, pages mobiles, and sends control messages before traffic assignment.
- Traffic channels: Carry user payload plus in-band signaling such as power-control or handover commands; include power-control and associated subchannels.
7.3 Reverse Link
- Uplink uses long-code phase offsets for user separation; Walsh codes provide 64-ary modulation rather than channelization.
- Access channel: Mobile-to-base signaling path before traffic assignment.
- Reverse traffic channel: Carries payload plus auxiliary services or signaling.
- Forward vs. reverse differences: Convolutional rates (1/2 vs. 1/3), roles of Walsh codes, long-code usage (scrambling vs. user ID), and modulation (QPSK vs. OQPSK).
7.4 Enhancements
- Voice codec: Qualcomm CELP (QCELP) with rate adaptation based on thresholds tied to ambient noise.
- RAKE receiver, power control, voice activity detection, and soft/hard handover (soft handover establishes the new link before releasing the old one).
7.5 IS-95 Features
- Large/soft capacity, soft handover, high voice quality, low transmit power, voice activity, and inherent secrecy.
8. 3G
8.1 Characteristics
- Global: IMT-2000 unifies multiple systems with common services and interworking for worldwide roaming.
- Integrated: Combines paging, cellular, and satellite systems.
- Multimedia: Supports high-quality voice, variable-rate data, and rich video.
- Personalized: Users leverage a unique personal telecom number (PTN) across terminals for personal mobility.
8.2 WCDMA
- Network: UE (ME + USIM), UTRAN (RNC + Node B), and core network (CN) for external connectivity and control.
- Channels: Logical channels describe payload types; MAC maps them to transport channels (dedicated DCH or shared channels such as RACH, BCH, PCH, FACH, CPCH, DSCH, HS-DSCH), which then map to physical channels (uplink/downlink).
- Air interface: 5 MHz bandwidth, 3.84 Mcps chip rate, AMR voice codec, supports synchronous/asynchronous base stations, combined inner/outer loop power control, downlink open/closed-loop transmit diversity, pilot-aided coherent demodulation, convolutional/Turbo coding, BPSK uplink and QPSK downlink.
8.3 CDMA2000
- Network: Similar layered structure (see diagram).
- Air interface: CDMA2000-compatible with IS-95; bandwidth \(N\times1.25\) MHz with chip rate \(N\times1.2288\) Mcps (N = 1,3,6,9,12), 8k/13k QCELP or 8k EVRC voice, GPS/GLONASS synchronization, dual-loop power control, downlink OTD/STS transmit diversity, pilot-aided coherent demodulation, convolutional/Turbo coding, BPSK uplink and QPSK downlink.
8.4 TD-SCDMA
- Network: See architecture figure.
- Key technologies: Mixes TDMA/CDMA/FDMA/SDMA for adaptive resource allocation; flexible uplink-downlink timeslot partitioning; power control to counter breathing/near-far; smart antennas to form user-specific beams and exploit spatial resources.
- Advantages: High spectral efficiency, flexible coverage, and compatibility with asymmetric traffic.
9. 4G
9.1 Characteristics
- All-IP design with high data rates, low latency, adaptive QoS, seamless mobility, and tight integration with Internet services.
9.2 LTE
- Key technologies: Supports FDD, TDD, and half-duplex FDD. Frame type 1 (10 ms, 20 slots, 10 subframes) and type 2 (two half-frames with 0.5 ms slots plus special downlink/uplink pilot/protection slots). Employs OFDM for spectral efficiency and multipath resilience.
- CDMA vs. OFDM comparison:
- Modulation: CDMA enforces uniform modulation per link; OFDM allows per-link adaptive modulation for better spectral-use/BEP trade-offs.
- PAPR: CDMA PAPR ≈ 5–11 dB and grows with rate/codes; OFDM’s non-constant envelope makes it sensitive to nonlinearity without mitigation.
- Narrowband interference: CDMA spreads interference over a small fraction of the spectrum; OFDM can disable interfered subcarriers or rely on FEC/lower-order modulation.
- Multipath: CDMA leverages RAKE but loses accuracy beyond ~7–8 significant paths; OFDM lowers symbol rate via serial-to-parallel conversion and adds CPs to suppress ISI, albeit with bandwidth/energy penalties.
- Equalization: CDMA often forgoes equalizers because spreading handles delay spread; OFDM already exploits frequency diversity, so extra equalization is usually unnecessary.
9.3 LTE-Advanced
- Key technologies: Uplink SC-FDMA enhancements, improved interference suppression, modulation up to 64QAM; carrier aggregation for wider effective bandwidth while staying backward compatible; relay nodes to extend coverage and edge throughput with lower complexity than base stations; MIMO upgrades (downlink 8×8, uplink 4×4); coordinated multi-point transmission, distributed antennas, inter-base-station coordination, and enhanced multimedia broadcast/multicast services.