π Lecture 20 β EDGE
π Overview: This lecture covers Enhanced Data rates for GSM Evolution (EDGE), a 2.5G technology bridging GPRS and 3G. It explains EDGE's modulation and coding schemes, link adaptation techniques, capacity planning strategies, and the Dynamic Abis pool concept for efficient resource allocation.
ποΈ Topics Covered
The lecture begins with Walsh Codes and the IS-95 Reverse Link as review, then introduces EDGE as an evolution of GPRS. It covers Modulation and Coding Schemes (MCS) with 8-PSK modulation, Link Adaptation and Incremental Redundancy for optimizing throughput, Capacity Planning with frequency reuse patterns, and the Dynamic Abis pool concept for efficient backhaul.
π Lecture Summary
Last Lecture Review
The review covers IS-136 and CDMA/IS-95 advantages including frequency diversity, multipath resistance, privacy, and graceful degradation. Drawbacks include self-jamming, the near-far problem, and soft handoff. The IS-95 forward channels are Pilot Channel, Sync Channel, Paging, and Traffic, while reverse channels are Access Channels and Traffic.
Walsh Codes
Walsh codes are orthogonal codes used in CDMA systems for channel separation. A 2x2 Walsh Matrix assigns User 1 the code (1, 1) and User 2 the code (1, -1). The lecture also references a 4x4 Walsh matrix for larger systems.
IS-95 Reverse Link
The reverse link consists of up to 94 logical channels each occupying the same bandwidth of 1228 KHz. It supports 32 access channels and 62 traffic channels. The access channel is used to initiate a call, to respond to the paging channel, and for location update. In the reverse link, the convolutional encoder has a rate of 1/3, thus tripling the effective rate to a maximum of 28.8 kbps.
The IS-95 CDMA Reverse Channel uses OQPSK for power efficiency while QPSK demodulation remains easy, operating in the 869-894 MHz range. There is no spreading of data using orthogonal codes; instead, data from the block interleaver is grouped in units of 6 bits that serve as an index to select a row of the 64x64 Walsh matrix, and that row substitutes for the input. This expands the data rate by a factor of 64/6 to 307.2 kbps.
Enhanced Data rates for GSM Evolution
GPRS data rates still fall short compared to 3G. The delay in 3G deployment led to EDGE. Phase 1 (Release β99 & 2002 deployment) supports best-effort packet data at speeds up to about 384 kbps. Phase 2 (Release β2000 & 2003 deployment) adds Voice over IP capability.
π Definition β EDGE: Enhanced Data rates for GSM Evolution, a 2.5G technology that provides higher data rates than GPRS using 8-PSK modulation.
GPRS Architecture
EDGE architecture is similar to GPRS, but with important changes for higher data rates, mainly the modulation scheme. GMSK is used in GPRS, transmitting only one bit per symbol. In EDGE, Octogonal PSK (8-PSK) is used, enabling a threefold higher data rate of 59.2 kbps per radio time slot, achieved by transmitting 3 bits per symbol. GMSK has constant amplitude modulation while 8-PSK has variations in amplitude, changing radio frequency characteristics and requiring changes in the Base Station. This leads to minor hardware and software changes but major network performance improvements.
π Formula: Peak data rate = 8 time slots Γ 59.2 kbps per slot = 473 kbps (theoretical maximum)
The lecture presents nine modulation and coding schemes (MCS) providing different throughput levels. Though GMSK is more robust, 8-PSK gives higher throughput. However, increased data rate comes at the price of decreased system sensitivity, impacting coverage and network planning. A key advantage in EDGE is that switching between different coding schemes happens easily β a data block can be sent with better protection on failure, unlike GPRS where retransmission uses the same protection as the initial transmission.
π‘ Why this matters: The ability to switch coding schemes dynamically means EDGE can adapt to changing radio conditions, maintaining throughput where GPRS would fail.
Link Adaptation and Incremental Redundancy
Link Adaptation (LA) adjusts the MCS as propagation conditions change, maximizing throughput per channel based on bit error probability (BEP) measurements. Incremental Redundancy (IR) improves throughput by automatically adapting transmitted redundancy to channel conditions, achieved through ARQ and FEC.
π Definition β Incremental Redundancy: A technique that sends redundancy only when necessary, using Generalized Type-II ARQ with finer granularity of code rate. If the first transmission fails, additional redundant bits are sent rather than the entire packet.
Capacity Planning in EDGE
Capacity planning is similar to GPRS, but higher throughput per radio time slot changes some aspects. The reuse pattern defines the number of cells in a cluster using different frequencies. A frequency reuse of 3/9 means each frequency is used once in three sites per cluster, where each site has three sectors. Reuse for control channels and traffic channels are independent. Typically, 4/12 is used for control and 1/3 for traffic, though other combinations are possible depending on performance requirements, environment, and spectrum availability. Higher frequency reuse yields higher throughput and less delay. Time slot capacities have a dynamic range depending on users.
1/3 Frequency Re-use (EDGE Compact) uses 3 Γ 200 kHz carriers reused in every site, requiring less than 1 MHz Γ 2 for initial deployment with 3 sectors per site. EDGE-capable and non-EDGE-capable TRX in one sector can be configured to have only one BCCH. TBF parameter setting makes it possible for TBFs of GPRS and EDGE radio networks to be multiplexed dynamically on one time slot, though this should be avoided as performance suffers in both uplink and downlink. In uplink, GPRS suffers due to large amounts of 8-PSK retransmissions. In downlink, it is due to GMSK modulation where 8-PSK can carry higher data rates for EDGE.
Dynamic Abis in EDGE
The Abis interface connects the Base Station (BS) and Base Station Controller (BSC). Voice signals are carried in 16 kbps Abis channels, which are static for GSM/GPRS. 8-PSK changes data rate from 8.8 kbps to 59.2 kbps, which is insufficient for data beyond MCS-2. Since this data rate is not always present, the Dynamic Abis concept is used in EDGE, where the BSC allocates Abis capacity from a Dynamic Abis Pool (DAP) for data calls from EGPRS when needed.
Benefits for operators include migration to wireless multimedia services, improved customer satisfaction, and early market deployment of third-generation-type applications. For users, benefits include improved quality of service, personal multimedia services, and potentially lower price per bit.
π‘ Why this matters: Dynamic Abis efficiently uses backhaul capacity, allocating high-bandwidth channels only when EDGE users actually need them, rather than reserving them statically.
β Key Takeaways
EDGE is a critical evolutionary step that extends GPRS with 8-PSK modulation, enabling 3 bits per symbol and a per-slot data rate of 59.2 kbps for a theoretical peak of 473 kbps. The nine MCS schemes allow dynamic tradeoffs between throughput and robustness, with Link Adaptation and Incremental Redundancy working together to maximize throughput under varying channel conditions. Capacity planning in EDGE uses frequency reuse patterns like 1/3 for traffic and 4/12 for control, with higher reuse yielding better throughput. The Dynamic Abis Pool efficiently allocates backhaul capacity on demand, solving the problem of variable data rates exceeding static channel capacities. Multiplexing GPRS and EDGE traffic on the same time slot should be avoided to prevent performance degradation from mixed modulation schemes.
π§ Quick Revision Questions
- How does 8-PSK modulation increase data rates compared to GMSK, and what is the resulting per-time-slot data rate?
- What is the difference between Link Adaptation and Incremental Redundancy in EDGE?
- Why should GPRS and EDGE traffic not be multiplexed on the same time slot?
- What frequency reuse patterns are typically used for control channels versus traffic channels in EDGE?
- How does the Dynamic Abis Pool (DAP) improve backhaul efficiency compared to static Abis allocation?
π Lecture 21 β WCDMA (Part I)
π Overview: This lecture introduces UMTS (Universal Mobile Telecommunications System) and its core WCDMA air interface. It covers UMTS service classes, UTRAN architecture, and the radio interface protocol architecture, explaining how 3G networks differ from GSM, particularly in handling data traffic, quality of service, and channel access using CDMA codes.
ποΈ Topics Covered
The lecture begins with a review of the last lecture (Walsh codes, IS-95 reverse link, EDGE). It then covers UMTS service classes including conversational, streaming, interactive, and background classes. UTRAN architecture is explained with Node B and RNC roles. The radio interface protocol architecture details layers 1, 2 (MAC and RLC), and 3 (RRC), plus PDCP and BMC. The protocol model for UTRAN is presented, followed by logical channels (BCCH, CCCH, CTCH, DCCH, DTCH, PCCH) in WCDMA.
π Lecture Summary
Last Lecture Review
The review covered Walsh Codes, IS-95 Reverse Link, EDGE Introduction, Modulation and Coding Schemes, Link Adaptation and Incremental Redundancy, Capacity Planning, and Dynamic Abis pool.
UMTS
UMTS networks have a predominance of data traffic unlike GSM networks. The data rate is significantly higher than GSM/GPRS/EDGE. Major changes from previous networks include:
- Max user data rate up to 384 kbps
- Efficient handover between different operators and technologies (e.g., GSM and UMTS)
- Ability to deliver at requested bandwidth
- Ability to deliver different services with the required quality
WCDMA Radio Fundamentals
WCDMA (Wideband Code Division Multiple Access) has emerged as the most preferred and adopted technology for 3G air interface. Major differences between WCDMA and GSM are:
- 5 MHz channel bandwidth compared to 200 KHz in GSM
- Packet data scheduling is load based, unlike time slot based in GSM
- Theoretically only one frequency channel, while GSM uses many channels
- Quality control is done using RRM algorithms, while it was done by frequency planning in GSM
- Users/cell/channel are separated by codes, unlike time or frequency in GSM
π‘ Why this matters: The shift from time/frequency division to code division fundamentally changes how capacity and interference are managed in 3G networks.
Service classes in UMTS
In 3G networks, Mobile Equipment (ME) can establish multiple connections simultaneously. The network allows efficient cooperation between applications with diverse Quality of Service (QoS) requirements.
Quality is defined by two main parameters:
- Guaranteed and max bit rate possible (kbps)
- Permissible delay (ms)
Based on QoS criteria, multimedia services are classified into:
π Definition β Conversational class: The most delay-sensitive applications, e.g., video telephony, VoIP.
π Definition β Streaming class: Flow which is steady and continuous, it is server to user.
π Definition β Interactive class: Web browsing is an example. A user may request timetables of buses, trains, or flight schedule.
π Definition β Background class: Short messages, file transfer, email that have the least stringent requirements of QoS.
UTRAN and System Architecture
UTRAN (UMTS Terrestrial Radio Access Network) provides the radio access network for UMTS.
USIM (Universal Subscriber Identity Module) contains authentication information, associated algorithms, encryption, and subscriber-related information.
ME (Mobile Equipment) is user independent.
Base Station (Node B) in WCDMA is more complex than BS in GSM. Its functions include:
- Handover channel management
- Baseband conversion
- Channel encoding and decoding
- Interfacing to other network elements
Radio Network Controller (RNC)
Concerning one connection between UTRAN and one UE, the following roles of RNCs exist:
- Serving RNC: Controls the connections to a UE
- Drift RNC: Lends its resources to the Serving RNC for a particular UE
Each RNC also has a Controlling RNC role towards its Node Bs.
Radio Interface protocol architecture
Layer 1 (Physical Layer):
- The actual medium of transfer
- Main functions include RF processing, modulation/demodulation, multiplexing/demultiplexing of physical channels
- Error detection and correction, rate matching, power control, synchronization
Layer 2 has two main sub-layers: RLC (Radio Link Control) and MAC (Medium Access Control).
MAC (Medium Access Control):
- Responsible for mapping logical channels to transport channels
- An interface between L1 & L3, provides packet multiplexing/demultiplexing
- Measurement related to traffic volume on logical channels and reporting to layer 3
RLC (Radio Link Control):
- Segmentation and reassembly of variable size data packets
- Error correction by retransmission and ACKed data transfer mode
- Controlling rate flow, concatenation, cyphering, and preservation of higher-order PDUs
- Operates in three modes (transparent, unacknowledged, acknowledged) as in GPRS
Layer 3 contains sub-layers; RRC (Radio Resource Control) interacts with layer 2:
- Handles control plane signaling between UE and network in connected mode
- Responsible for bearer functions like establishment, release, maintenance, and reconfiguration in the user plane and of radio resources in control plane
- Functions include radio resource management and mobility management, as well as power control, routing, and paging
Packet Data Convergence Protocol (PDCP):
- Major functions: compression of PDU at transmitting end and decompression at receiving end in all three modes of RLC
Broadcast β Multicast Control (BMC):
- Functions only in transparent and unacknowledged modes
- Provides broadcast/multicast scheduling and transmission of user data
Protocol Model for UTRAN
The UTRAN protocol structure is based on a model with horizontal layers (Radio Network Layer, Transport Network Layer) and vertical planes (Control Plane, User Plane, Transport Network Control Plane). The Control Plane handles signaling, while the User Plane carries user data.
Logical Channels in WCDMA
| Channel | Abbreviation | Functionality |
|---|---|---|
| Broadcast Common Control channel (DL) | BCCH | Transmits the system control information |
| Common Control channel (UL/DL) | CCCH | Used (usually by UE) for transmitting info related to control between network and UE |
| Common Traffic Channel (DL) | CTCH | Used to transmit dedicated info to a group of UEs |
| Dedicated Control Channel (UL/DL) | DCCH | Dedicated channel for control related information between UEs and network |
| Dedicated Traffic Channel (UL/DL) | DTCH | Similar to DCCH except that it is used for user information |
| Paging Control channel (DL) | PCCH | Used to page info the UE |
β Key Takeaways
UMTS service classes (conversational, streaming, interactive, background) are defined by delay sensitivity and bit rate requirements, with conversational being most delay-sensitive and background least stringent. The WCDMA air interface uses 5 MHz channels, employs code division for user separation, and relies on RRM algorithms for quality control instead of frequency planning. UTRAN architecture includes Node B (base station with complex encoding/handover functions) and RNC (with Serving, Drift, and Controlling roles). The radio interface protocol has Layer 1 (physical), Layer 2 (MAC and RLC sublayers), and Layer 3 (RRC), with additional PDCP for compression and BMC for broadcast/multicast. Logical channels such as BCCH, CCCH, DCCH, DTCH, and PCCH separate control and traffic information in the WCDMA system.
π§ Quick Revision Questions
- What are the four service classes in UMTS and which one is the most delay-sensitive?
- What is the channel bandwidth in WCDMA compared to GSM, and how are users separated in WCDMA?
- What are the three roles of a Radio Network Controller (RNC) and what does each do?
- What are the main functions of the RLC sublayer in the radio interface protocol architecture?
- List three logical channels in WCDMA and explain the purpose of each.
π Lecture 22 β WCDMA (Part II)
π Overview: This lecture continues the exploration of WCDMA technology within UMTS networks, focusing on the physical layer procedures that enable reliable communication. It covers how data is spread and scrambled, the different types of channels used, and essential signaling operations like random access, cell searching, and power control, all of which are critical for network efficiency and mobile device functionality.
ποΈ Topics Covered
The lecture begins with a review of UMTS architecture and protocol layers, then dives into physical layer procedures including spreading and scrambling with channelization (OVSF codes) and scrambling (Gold/S(2) codes). It explains the three channel concepts (logical, transport, physical) and maps transport channels to physical channels, with a focus on the uplink dedicated channel (DPDCH and DPCCH). Finally, it covers signaling procedures: RACH operation, the three-step cell searching process using synchronization channels, and the three types of power control (open loop, fast closed loop, and outer loop).
π Lecture Summary
Last Lecture Review
The lecture opens with a review of UMTS, its service classes, UTRAN architecture, radio interface protocol architecture, protocol models, and logical channels. Air interface parameters are specified: a carrier spacing of 5 MHz, a chip rate of 3.84 Mcps, a frame length of 10 ms (38400 chips), 15 slots per frame (2560 chips/slot), and spreading factors (SF) ranging from 4 to 256 uplink and 4 to 512 downlink, supporting channel rates from 7.5 Kbps to 960 Kbps.
Spreading and Scrambling
Spreading increases the signal bandwidth and includes two distinct operations. Channelization increases the signal bandwidth using orthogonal codes. Scrambling does not affect the signal bandwidth; it uses pseudo-noise codes.
Channelization codes are orthogonal codes based on the Orthogonal Variable Spreading Factor (OVSF) technique. These codes are fully orthogonal (they do not interfere) only if they are time-synchronized. They can separate transmissions from a single source; in the downlink, they separate different users within one cell/sector. Since limited orthogonal codes must be reused in every cell, this creates an interference problem if two cells use the same code. The solution is to use scrambling codes to reduce inter-base-station interference. It is possible for two mobiles to use the same channelization codes. To separate different users in the uplink, scrambling codes are used. One code tree is used with one scrambling code on top of the tree.
Scrambling involves multiplying the code sequence with a pseudorandom scrambling code. This code can be a long code (a Gold code with a 10 ms period) or a short code (S(2) code). In the downlink, scrambling codes reduce inter-base-station interference. Typically, each Node B has only one scrambling code for UEs to separate base stations. Since a code tree under one scrambling code is used by all users in its cell, proper code management is needed.
Channel Concept
Three separate channel concepts exist in UTRA: logical, transport, and physical channels.
- Logical channels define what type of data is transferred.
- Transport channels define how and with which type of characteristics the data is transferred by the physical layer.
- Physical channels define the exact physical characteristics of the radio channel.
Transport Channels -> Physical Channels
Transport channels contain data generated at higher layers, which is carried over the air and mapped to different physical channels. Data is sent in transport blocks from the MAC layer to the physical layer every 10 ms. The transport format is identified by the Transport Format Indicator (TFI). Several transport channels can be multiplexed to form a single Coded Composite Transport Channel (CCTrCh). The physical layer combines TFI information into the Transport Format Combination Indicator (TFCI), which indicates which transport channels are active for the current frame.
There are two types of transport channels:
- Dedicated channels: Reserved for a single user, support fast power control and soft handover.
- Common channels: Can be used by any user at any time, do not support soft handover but some support fast power control.
In addition, there are physical channels for signaling purposes. The mapping is as follows:
- DCH (UL/DL Dedicated channel) maps to DPDCH (Dedicated Physical Data Channel) and DPCCH (Dedicated Physical Control Channel).
- RACH (UL Random Access Channel) maps to PRACH (Physical Random Access Channel).
- CPCH (UL Common Packet Channel) maps to PCPCH.
- BCH (DL Broadcast Channel) maps to P-CCPCH (Primary Common Control Physical Channel).
- FACH and PCH (DL) map to S-CCPCH (Secondary Common Control Physical Channel).
- DSCH (DL Downlink Shared Channel) maps to PDSCH.
- Signaling channels include SCH (Synchronization Channel), CPICH (Common Pilot Channel), AICH (Acquisition Indication Channel), PICH (Paging Indication Channel), CSICH, and CD/CS-ICH.
UL Dedicated Channel DCH
To overcome discontinuous transmission (DTX) and its audible interference to audio equipment, two dedicated physical channels are used: the Dedicated Physical Control Channel (DPCCH) and the Dedicated Physical Data Channel (DPDCH). They use code multiplexing instead of time multiplexing.
The DPCCH has a fixed spreading factor of 256 and carries physical layer control information with four fields:
- Pilot: Used for channel estimation and SIR estimate for power control.
- TFCI: Indicates the bit rate, channel decoding, and interleaving parameters for every DPDCH frame.
- FBI (Feedback Information): Used for transmission diversity in the downlink.
- TPC (Transmission Power Control): Contains the power control command.
The DPDCH has a spreading factor from 4 to 256 and its data rate may vary on a frame-by-frame basis, as informed on the DPCCH channel. Parallel channel codes can provide 2 Mbps user data.
UL receiver in BS
The receiver in the base station performs these steps:
- Starts receiving the frame, de-spreading the DPCCH and buffering the DPDCH according to the max bit rate.
- For every slot: Obtains a channel estimate using pilot bits and estimates SIR, sends a TPC command in the downlink to control UE uplink power, and decodes the TPC bit to adjust downlink power.
- For every 2nd or 4th slot: Decodes FBI bits to adjust antenna phases and amplitude for transmission diversity.
- For every 10 ms frame: Decodes TFCI information to obtain the bit rate.
- For each transmission time interval (TTI) of 10, 20, 40, or 80 ms: Decodes the DPDCH data.
RACH Operation
The Random Access Channel (RACH) operation is a signaling procedure:
- The UE first sends a preamble.
- The preamble's SF is 256 and contains a signature sequence of 16 symbols, for a total length of 4096 chips.
- The UE waits for an acknowledgement via the Acquisition Indication Channel (AICH) from the BS.
- If no AICH is received after a period, the UE sends another preamble with higher power.
- When the AICH is received, the UE sends a 10 or 20 ms message part.
- The SF for the message part is from 32 to 256.
Synchronisation Channel (SCH) β Cell Searching
Cell searching using the SCH is a three-step procedure for a UE to find and synchronize with a cell:
- Slot Synchronisation: The UE searches for the 256-chip primary synchronisation code, which is common to all cells and the same in every slot. Detecting peaks in the filter output corresponds to finding the slot boundary.
- Frame Synchronisation: The UE seeks the largest peak from the secondary synchronisation code (SSC). There are 64 unique SSC sequences, each with 15 SSCs. The UE must know 15 successive SSCs from the S-SCH to determine the code group and find the frame boundary.
- Scrambling Code Identification: Each code group has 8 primary scrambling codes. The correct one is found by testing each possible scrambling code over the Common Pilot Channel (CPICH) of that cell.
Power Control
Three types of power control are detailed:
- Fast Closed Loop PC β Inner Loop PC: Uses feedback information. Uplink PC solves the near-far problem. Downlink PC ensures enough power for mobiles at the cell edge. There is one PC command per slot, giving a rate of 1500 Hz. Two special cases exist:
- Soft handover: The mobile receives multiple power control commands. A βpower downβ command has higher priority over a βpower upβ command.
- Compressed mode: A large step size is used after a compressed frame to allow the power level to converge more quickly to the correct value after a break.
- Closed Loop PC - Outer Loop PC: This sets the SIR target to maintain a certain frame error rate (FER). It is operated at the radio network controller (RNC).
- Open Loop PC: This uses no feedback information.
β Key Takeaways
The most critical concepts from this lecture are the distinction between channelization (OVSF codes for user separation within a cell) and scrambling (for base station separation), the three-layer channel model (logical, transport, physical) and their specific mapping (especially DCH to DPDCH/DPCCH with DPCCH's four control fields), and the three physical layer procedures: the preamble/AICH-based RACH operation, the three-step cell search (slot, frame, scrambling code), and the three power control loops (open, fast closed inner, and slow outer). Understanding the UL receiver's slot-by-slot and frame-by-frame tasks is also essential.
π§ Quick Revision Questions
- What is the purpose of scrambling codes in the WCDMA downlink, and how do they solve a problem created by channelization codes?
- What are the four fields carried by the Uplink Dedicated Physical Control Channel (DPCCH), and what is the function of each?
- Describe the three steps a UE performs to find and synchronize with a cell using the Synchronization Channel (SCH).
- In the context of the RACH operation, what happens if a UE does not receive an Acquisition Indication Channel (AICH) acknowledgement?
- Explain the difference between fast closed-loop power control and outer-loop power control, including where each is implemented and what parameter each one aims to adjust.