WCDMA HSPA and advanced receivers Timo Nihtil Ph

WCDMA, HSPA and advanced receivers Timo Nihtilä, Ph. Lic. (Ph. D. def. ) Senior Research Scientist Magister Solutions Ltd. 1

Readings related to the subject • General readings – WCDMA for UMTS – Harri Holma, Antti Toskala – HSDPA/HSUPA for UMTS – Harri Holma, Antti Toskala • Network planning oriented – Radio Network Planning and Optimisation for UMTS – Janna Laiho, Achim Wacker, Tomás Novosad – UMTS Radio Network Planning, Optimization and Qo. S Management For Practical Engineering Tasks – Jukka Lempiäinen, Matti Manninen 2 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Outline • Background • Key concepts – Code multiplexing – Spreading • Introduction to Wideband Code Division Multiple Access (WCDMA) • WCDMA Performance Enhancements – High Speed Packet Access (HSDPA/HSUPA) – Advanced features for HSDPA 3 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Background • Why new radio access system • Frequency Allocations • Standardization • WCDMA background and evolution • Evolution of Mobile standards • Current WCDMA markets 4

Why new radio access system • Need for universal standard (Universal Mobile Telecommunication System) • Support for packet data services – IP data in core network – Wireless IP • New services in mobile multimedia need faster data transmission and flexible utilization of the spectrum • FDMA and TDMA are not efficient enough – TDMA wastes time resources – FDMA wastes frequency resources • CDMA can exploit the whole bandwidth constantly • Wideband CDMA was selected for a radio access system for UMTS (1997) – (Actually the superiority of OFDM was not fully understood by then) 5 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Frequency allocations for UMTS • Frequency plans of Europe, Japan and Korea are harmonized • US plan is incompatible, the spectrum reserved for 3 G elsewhere is currently used for the US 2 G standards • IMT-2000 band in Europe: – FDD 2 x 60 MHz Expected air interfaces and spectrums, source: “WCDMA for UMTS” 6 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Standardization • WCDMA was studied in various research programs in the industry and universities • WCDMA was chosen besides ETSI also in other forums like ARIB (Japan) as 3 G technology in late 1997/early 1998. • During 1998 parallel work proceeded in ETSI and ARIB (mainly), with commonalities but also differences – Work was also on-going in USA and Korea 7 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Standardization • At end of 1998 different standardization organizations got together and created 3 GPP, 3 rd Generation Partnership Project. – 5 Founding members: ETSI, ARIB+TTC (Japan), TTA (Korea), T 1 P 1 (USA) – CWTS (China) joined later. • Different companies are members through their respective standardization organization. 8 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

WCDMA Background and Evolution • First major milestone was Release ‘ 99, 12/99 – Full set of specifications by 3 GPP – Targeted mainly on access part of the network • Release 4, 03/01 – Core network was extended – markets jumped over Rel 4 • Release 5, 03/02 – High Speed Downlink Packet Access (HSDPA) • Release 6, end of 04/beginning of 05 – High Speed Uplink Packet Access (HSUPA) • Release 7, 06/07 – Continuous Packet connectivity (improvement for e. g. Vo. IP), advanced features for HSDPA (MIMO, higher order modulation) 9 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

WCDMA Background and Evolution 3 GPP Rel -99 12/99 2001 2000 Japan 10 © Timo Nihtilä 3 GPP Rel 4 03/01 3 GPP Rel 5 (HSDPA) 03/02 2002 Europe (pre-commercial) TLT-5606 Spread Spectrum Techniques / 25. 4. 2008 2003 Europe (commercial) 3 GPP Rel 6 (HSUPA) 2 H/04 2005 HSDPA (commercial) 3 GPP Rel 7 HSPA+ 06/07 2006 Further Releases 2007 HSUPA (commercial)

Evolution of Mobile standards EDGE GSM HSCSD WCDMA FDD HSDPA/ HSUPA GPRS LTE TD-CDMA TDD HCR HSDPA/ HSUPA TD-SCDMA TDD LCR cdma 2000 1 XEV - DO cdma. One (IS-95) cdma 2000 1 XEV - DV 11 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Current WCDMA markets • Graph of the technologies adopted by the wireless users worldwide: • Over 3. 5 billion wireless users worldwide • GSM+WCDMA share currently over 88 % (www. umts-forum. org) • CDMA share is decreasing every year 12 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Million subscriber Current WCDMA markets • Over 200 million WCDMA subscribers globally (04/08) (www. umts-forum. org) – 10 % HSDPA/HSUPA users • Number of subscribers is constantly increasing 13 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Key concepts • CDMA • Spread Spectrum • Direct Sequence spreading • Spreading and Processing gain 14

Multiple Access Schemes FDMA TDMA Code Time 1 2 … N Frequency • Frequency Division Multiple Access (FDMA), different frequencies for different users – example Nordic Mobile Terminal (NMT) systems • Time Division Multiple Access (TDMA), same frequency but different timeslots for different users, – example Global System for Mobile Communication (GSM) – GSM also uses FDMA • Code Division Multiple Access (CDMA), same frequency and time but users are separated from each other with orthogonal codes 15 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Spread Spectrum • Means that the transmission bandwidth is much larger than the information bandwidth i. e. transmitted signal is spread to a wider bandwidth – Bandwidth is not dependent on the information signal • Benefits – More secure communication – Reduces the impact of interference (and jamming) due to processing gain • Classification – Direct Sequence (spreading with pseudo noise (PN) sequence) – Frequency hopping (rapidly changing frequency) – Time Hopping (large frequency, short transmission bursts) • Direct Sequence is currently commercially most viable 16 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Spread Spectrum • Where does spread spectrum come from – – – – 17 First publications, late 40 s First applications: Military from the 50 s Rake receiver patent 1956 Cellular applications proposed late 70 s Investigations for cellular use 80 s IS-95 standard 1993 (2 G) 1997/1998 3 G technology choice 2001/2002 Commercial launch of WCDMA technology © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Direct Sequence • In direct sequence (DS) user bits are coded with unique binary sequence i. e. with spreading/channelization code – The bits of the channelization code are called chips – Chip rate (W) is typically much higher than bit rate (R) – Codes need to be in some respect orthogonal to each other (cocktail party effect) • Length of a channelization code – defines how many chips are used to spread a single information bit and thus determines the end bit rate – Shorter code equals to higher bit rate but better Signal to Interference and Noise Ratio (SINR) is required • Also the shorter the code, the fewer number of codes are available – Different bit rates have different geographical areas covered based on the interference levels 18 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Direct Sequence • Transmission (Tx) side with DS – Information signal is multiplied with channelization code => spread signal • Receiving (Rx) side with DS – Spread signal is multiplied with channelization code – Multiplied signal (spread signal x code) is then integrated (i. e. summed together) • If the integration results in adequately high (or low) values, the signal is meant for the receiver 19 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Direct Sequence 20 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Direct Sequence 21 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Transmitted signal before spreading Power density (Watts/Hz) Processing gain and Spreading Despread narrowband signal Spread wideband signal Received signal before despreading Power density (Watts/Hz) R Frequency W Interference for the part we are interested in Frequency 22 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Received signal after despreading but before filtering Power density (Watts/Hz) Processing gain and Spreading Transmitted signal Interference Received signal after despreading and after filtering Power density (Watts/Hz) Frequency 23 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Processing gain and Spreading • Spread spectrum systems reduce the effect of interference due to processing gain • Processing gain is generally defined as follows: – G[d. B]=10*log 10(W/R), where ’W’ is the chip rate and ’R’ is the user bit rate • The number of users takes negative effect on the processing gain. The loss is defined as: – Lp = 10*log 10 k, where ’k’ is the amount of users • Processing gain when the processing loss is taken into account is – Gtot=10*log 10(W/k. R) • High bit rate means lower processing gain and higher power OR smaller coverage • The processing gain is different for different services over 3 G mobile network (voice, web browsing, videophone) due to different bit rates – Thus, the coverage area and capacity might be different for different services depending on the radio network planning issues 24 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Processing gain and Spreading • Processing gain is what gives CDMA systems the robustness against self-interference that is necessary in order to reuse the available 5 MHz carrier frequency over geographically close distances. • Examples: Speech service with a bit rate of 12. 2 kbps – processing gain 10 log 10(3. 84 e 6/12. 2 e 3) = 25 d. B – For speech service the required SINR is typically in the order of 5. 0 d. B, so the required wideband signal-to-interference ratio (also called “carrier-tointerference ratio, C/I ) is therefore “ 5. 0 d. B minus the processing” = -20. 0 d. B. – In other words, the signal power can be 20 d. B under the interference or thermal noise power, and the WCDMA receiver can still detect the signal. – Notice: in GSM, a good quality speech connection requires C/I = 9– 12 d. B. 25 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Introduction to Wideband Code Division Multiple Access (WCDMA) • Overview • Codes in WCDMA • Qo. S support • Network Architecture • Radio propagation and fading • RAKE receiver • Power Control in WCDMA • Diversity • Capacity and coverage 26

WCDMA System • WCDMA is the most common radio interface for UMTS systems • Wide bandwidth, 3. 84 Mcps (Megachips per second) – Maps to 5 MHz due to pulse shaping and small guard bands between the carriers • Users share the same 5 MHz frequency band time – UL and DL have separate 5 MHz frequency bands • High bit rates – With Release ’ 99 theoretically 2 Mbps both UL and DL – 384 kbps highest implemented • Fast power control (PC) => Reduces the impact of channel fading and minimizes the interference 27 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

WCDMA System • Soft handover – Improves coverage, decreases interference • Robust and low complexity RAKE receiver – Introduces multipath diversity • Variable spreading factor – Support for flexible bit rates • Multiplexing of different services on a single physical connection – Simultaneous support of services with different Qo. S requirements: • real-time – E. g. voice, video telephony • streaming – streaming video and audio • interactive – web-browsing • background – e-mail download 28 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Codes in WCDMA • Channelization Codes (=short code) – Codes from different branches of the code tree are orthogonal – Length is dependent on the spreading factor – Used for • channel separation from the single source in downlink • separation of data and control channels from each other in the uplink – Same channelization codes in every cell / mobiles and therefore the additional scrambling code is needed • Scrambling codes (=long code) – – – Very long (38400 chips = 10 ms =1 radio frame), many codes available Does not spread the signal Uplink: to separate different mobiles Downlink: to separate different cells The correlation between two codes (two mobiles/Node. Bs) is low • Not fully orthogonal 29 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Codes in WCDMA • For instance, the relation between downlink physical layer bit rates and codes Symbol_rate = Chip_rate/SF Spreading Factor (SF) 512 256 128 64 32 16 8 4 4, with 3 parallel codes 30 © Timo Nihtilä Bit_rate = Symbol_rate*2 Channel symbol rate (ksps) 7. 5 15 30 60 120 240 480 960 2880 Control channel (DPCCH) overhead Channel bit rate (kbps) 15 30 60 120 240 480 960 1920 5760 TLT-5606 Spread Spectrum Techniques / 25. 4. 2008 User bit rate with coding = Channel_bit_rate/2 DPDCH channel bit rate range (kbps) 3– 6 12– 24 42– 51 90 210 432 912 1872 5616 Maximum user data rate with ½rate coding (approx. ) 1– 3 kbps Half rate speech 6– 12 kbps Full rate speech 20– 24 kbps 45 kbps 105 kbps 144 kbps 215 kbps 456 kbps 384 kbps 936 kbps 2. 3 Mbps 2 Mbps

Qo. S Support • Key Factors: – Simultaneous support of services with different Qo. S requirements: • up to 210 Transport Format Combinations, selectable individually for every radio frame (10 ms) • going towards IP core networks greatly increases the usage of simultaneous applications requiring different quality, e. g. real time vs. non-real time – Optimized usage of different transport channels for supporting different Qo. S 31 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Qo. S support Example: Downlink Shared Channel Data Rate 2 Mbps 10 ms USER 1 USER 2 USER 3 USER 1 USER 2 Code 5 USER 4. . Downlink Dedicated Channels USER 4 Code 4 USER 3 Code 3 USER 2 Code 2 USER 1 Code 1 Time 32 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

UMTS Terrestrial Radio Access Network (UTRAN) Architecture • New Radio Access network needed mainly due to new radio access technology Uu interface RNC • Core Network (CN) is based on GSM/GPRS • Radio Network Controller (RNC) corresponds roughly to the Base Station Controller (BSC) in GSM • Node B corresponds roughly to the Base Station in GSM – Term “Node B” is a relic from the first 3 GPP releases 33 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008 Iub interface Node. B UE CN Node. B Iur interface UE Node. B RNC UTRAN

UMTS Terrestrial Radio Access Network (UTRAN) Architecture • Radio network controller (RNC) – Owns and controls the radio resources in its domain – Radio resource management (RRM) tasks include e. g. the following • • • 34 Mapping of Qo. S Parameters into the air interface Air interface scheduling Handover control Outer loop power control Call Admission Control Setting of initial powers and SIR targets Radio resource reservation Code allocation Load Control © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

UMTS Terrestrial Radio Access Network (UTRAN) Architecture • Node B – Main function to convert the data flow between Uu and Iub interfaces – Some RRM tasks: • Measurements • Inner loop power control 35 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Radio propagation and fading • A transmitted radio signal goes through several changes while traveling via air interface to the receiver – reflections, diffractions, phase shifts and attenuation • Due to length difference of the signal paths, multipath components of the signal arrive at different times to the receiver and can be combined either destructively or constructively – Depends on the phases of the multipath components 36 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Radio propagation and fading • Example of the fast fading channel of a function of time • Opposite phases of two random multipath components arriving at the same time cancel each other out – Results in a fade • Coherent phases are combined constructively 37 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

RAKE receiver • Every multipath component arriving at the receiver more than one chip time (0. 26 μs) apart can be distinguished by the RAKE receiver – 0. 26 μs corresponds to 78 m in path length difference • RAKE assigns a “finger” to each received component (tap) and alters their phases based on a channel estimate so that the components can be combined constructively Transmitted symbol Received symbol at each time slot Finger #1 Finger #2 Finger #3 38 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008 Phase modified using the channel estimate Combined symbol

Power Control in WCDMA • The purpose of power control (PC) is to ensure that each user receives and transmits just enough energy to have service but to prevent: – Blocking of distant users (near-far-effect) – Exceeding reasonable interference levels Without PC received power levels would be unequal UE 1 UE 2 UE 3 UE 2 UE 1 UE 2 UE 3 39 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008 With ideal PC received power levels are equal

Power Control in WCDMA 1. Open loop power control • • Only for the initial power setting of the MS Based on distance attenuation estimation from the downlink pilot signal 2. Inner loop transmitter power control (CL TPC) at a rate of 1500 Hz • • Mitigates fading processes (fast and slow fading) Tx power is adjusted up/down to reach SIR target Both in UL and DL Uses quality targets in MS / BS 3. Outer loop PC at the rate of 100 Hz • • 40 Sets the quality target used by the inner loop PC Compensates the changes in the propagation conditions Adjusts the quality target Both in UL and DL © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Power Control in WCDMA • Inner loop power control in the uplink Outer loop PC (running in the radio network controller, RNC) defines SIR target for the BS. – If the measured SIR at BS is lower than the SIR-target, the MS is commanded to increases its transmit power. Otherwise MS is commanded to decrease its power – Power control dynamics at the MS is 70 d. B – 41 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Power Control in WCDMA • Inner loop power control in downlink: – – – 42 Outer loop PC (running in the MS) defines SIR target for the MS If the measured SIR at the MS is lower than the SIR-target, the BS is commanded to increases its transmit power for that MS. Otherwise, BS is commanded to decrease its power. Power control rate 1500 Hz Power control dynamics is dependent on the service There’s no near-far problem in DL due to one-to-many scenario. However, it is desirable to provide a marginal amount of additional power to mobile stations at the cell edge, as they suffer from increased other-cell interference. © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Power Control in WCDMA • Example of inner loop power control behavior: • With higher velocities channel fading is more rapid and 1500 Hz power control may not be sufficient 43 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Power Control in WCDMA • Inner loop power control tries to keep the received SIR as close to the target SIR as possible. • However, the constant SIR alone does not actually guarantee the required frame error rate (FER) which can be considered as the quality criteria of the link/service. There’s no unique SIR that automatically gives a certain FER – FER is a function of SIR, but also depends on mobility and propagation environment. – • 44 Therefore, the frame reliability information has to be delivered to outer loop control, which can tune the SIR target if necessary. © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Diversity • Transmitting on a single path only can lead to serious performance degradation due to fading • As fading is independent between different times and spaces it is reasonable to use the available diversity of them to decrease the probability of a deep fade – The more there are paths to choose from, the less likely it is that all of them have a poor energy level • There exists different types of diversity which can be used to improve the quality, e. g. : – Multipath • RAKE receiver exploits taps arriving at different times – Macro • Different Node Bs send the same information – Site Selection Transmit Diversity (SSTD) • Maintain a list of available base stations and choose the best one, from which the transmission is received and tell the others not to transmit 45 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Diversity – Time • Same information is transmitted in different times – Receive antenna • Transmission is received with multiple antennas • Power gain and diversity gain – Transmit antenna • Transmission is sent with multiple antennas 46 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

WCDMA Handovers • WCDMA handovers can be categorized into three different types • Intra-frequency handover – WCDMA handover within the same frequency and system. Soft, softer and hard handover supported • Inter-frequency handover – Handover between different frequencies (carriers) but within the same system – E. g. from one WCDMA operator to another – Only hard handover supported • Inter-system handover – Handover between WCDMA and another system, e. g. from WCDMA to GSM – Only hard handover supported 47 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

WCDMA Handovers • Soft handover – Handover between different Node Bs – Several Node Bs transmit the same signal to the UE which combines the transmissions • Advantages: lower Tx power needed for each Node B and UE – lower interference, battery saving for UE • Disadvantage: resources (code, power) need to be reserved for the UE in each Node B – Excess soft handovers limit the capacity – No interruption in data transmission – Needs RNC duplicating frame transmissions to two Node Bs 48 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

WCDMA Handovers • Softer handover – Handover between two sectors of the same Node B • Special case of a soft handover • No need for duplicate frames • Hard handover – The source is released first and then new one is added – Short interruption in data flow 49 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

WCDMA Handovers • Some terminology – Active set (AS), represents the Node Bs to which the UE is in soft handover – Neighbor set (NS), represents the links that UE monitors but which are not already in active set Triggering time_1 Triggering time_2 BS 1 Received signal strength Threshold_1 Threshold_2 BS 1 dropped from the AS BS 2 from the NS reaches the threshold to be added to the AS 50 © Timo Nihtilä BS 2 is still after the triggering time above threshold and thus added to the AS TLT-5606 Spread Spectrum Techniques / 25. 4. 2008 BS 1 from the AS reaches the threshold to be dropped from the AS

Capacity and coverage • In WCDMA coverage and capacity are tight together: When the load increases, the interference levels increases, too, and therefore also increased transmit powers are needed in order to keep constant quality. – Due to finite power resources, the more users Node B serves the less power it has for each UE coverage will decrease – This leads to cell breathing: the coverage area changes as the load of the cell changes. • Therefore, the coverage and the capacity have to be planned simultaneously • • Radio resource management (RRM) is needed in WCDMA to effectively control cell breathing. 51 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Capacity and coverage • Received power of one user as a function of users per cell • Due to finite maximum Tx power of the UE coverage is usually limited by the uplink • Node B does not have this problem – There is enough Tx power to transmit very far to a single user if necessary – However, downlink Tx power is divided between all users and thus capacity is limited by the downlink 52 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

WCDMA evolution • High Speed Downlink Packet Access (HSDPA) • High Speed Uplink Packet Access (HSUPA) • Advanced receivers with HSDPA • Advanced HSDPA scheduling • Femto cells with HSDPA 53

High Speed Downlink Packet Access (HSDPA) • The High Speed Downlink Packet Access (HSDPA) concept was added to Release 5 to support higher downlink data rates • It is mainly intended for non-real time traffic, but can also be used for traffic with tighter delay requirements. • Peak data rates up to 10 Mbit/s (theoretical data rate 14. 4 Mbit/s) • Reduced retransmission delays • Improved Qo. S control (Node B based packet scheduler) • Spectrally and code efficient solution 54 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

HSDPA features • Agreed features in Release 5 – Adaptive Modulation and Coding (AMC) • QPSK or 16 QAM – Multicode operation • Support of 1 -15 code channels (SF=16) – Short frame size (TTI = 2 ms) – Fast retransmissions using Hybrid Automatic Repeat Request (HARQ) • Chase Combining • Incremental Redundancy – Fast packet scheduling at Node B • E. g. Round robin, Proportional fair • Features agreed in Release 7 – Higher order modulation (64 QAM) – Multiple Input Multiple Output (MIMO) 55 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

HSDPA - general principle Channel quality (CQI, Ack/Nack, TPC) Data UE New base station functions • HARQ retransmissions • Modulation/coding selection • Packet data scheduling (short TTI) Users may be time and/or code multiplexed • Fast scheduling is done directly in Node-B based on feedback information from UE and knowledge of current traffic state. 56 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

HSDPA functionality • Scheduling responsibility has been moved from RNC to Node B • Due to this and the short TTI length (2 ms) the scheduling is dynamic and fast • Support for several parallel transmissions – When packet A is sent it starts to wait for an acknowledgement from the receiver, during which other packets can be sent via a parallel SAW (stopand-wait) channels Pkt A Pkt B Pkt C Pkt D Pkt E Pkt F 57 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008 Ack B

HSDPA functionality • UE informs the Node B regularly of its channel quality by CQI messages (Channel Quality Indicator) 58 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

HSDPA functionality • Node B can use channel state information for several purposes – In transport format (TFRC) selection • Modulation and coding scheme – Scheduling decisions • Non-blind scheduling algorithms can be utilized – HS-SCCH power control 59 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

HSDPA channels • User data is sent on High Speed Downlink Shared Channel (HSDSCH) • Control information is sent on High Speed Common Control Channel (HS-SCCH) • HS-SCCH is sent two slot before HS-DSCH to inform the scheduled UE of the transport format of the incoming transmission on HS-DSCH 60 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

High Speed Uplink Packet Access (HSUPA) • Peak data rates increased to significantly higher than 2 Mbps; Theoretically reaching 5. 8 Mbps • Packet data throughput increased, though not as high throughput as with HSDPA • Reduced delay from retransmissions. • Solutions – Layer 1 hybrid ARQ – Node. B based scheduling for uplink – Frame sizes 2 ms & 10 ms • Schedule in 3 GPP – Part of Release 6 – First specifications version completed 12/04 – In 3 GPP specs with the name Enhanced uplink DCH (E-DCH) 61 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

HSPA Peak Data Rates Downlink HSDPA Uplink HSUPA • Theoretical up to 14. 4 Mbps • Theoretical up to 5. 76 Mbps • Initial capability 1. 8 – 3. 6 Mbps • Initial capability 1. 46 Mbps # of codes Modulation 62 Max data rate # of codes TTI Max data rate 5 codes QPSK 1. 8 Mbps 2 x SF 4 2 ms 10 ms 1. 46 Mbps 5 codes 16 -QAM 3. 6 Mbps 2 x SF 2 10 ms 2. 0 Mbps 10 codes 16 -QAM 7. 2 Mbps 2 x SF 2 2 ms 2. 9 Mbps 15 codes 16 -QAM 10. 1 Mbps 2 x SF 2 + 2 x SF 4 2 ms 5. 76 Mbps 15 codes 16 -QAM 14. 4 Mbps © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Performance of advanced HSDPA features 63

Advanced receivers with HSDPA • UE receiver experiences significant interference from different sources – In a reflective environment the signal interferes itself – Neigboring base station signals interfere each other – One solution to decrease mainly own base station signal interference is to use an equalizer before despreading Own cell interference Other cell interference Own signal 64 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Advanced receivers with HSDPA • In a frequency-selective channel there is a significant amount of interfering multipaths • Linear Minimum Mean Squared Error (LMMSE) equalizer can be used to make an estimate of the original transmitted chip sequence before despreading – The interfering multipath components are removed – The channel becomes flat again 65 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Advanced receivers with HSDPA • LMMSE equalizer (Equ in the figure) offers a very good performance for the user especially near the base station • Using antenna diversity (1 x 2) the throughput can be doubled compared to a single antenna • Both techniques increase the cost of a mobile unit 66 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Advanced HSDPA scheduling • Node B has a limited amount of scheduling opportunities • The amount of data transmitted by the network must be maximized whilst offering the best possible quality of service to all users – The scheduling can be improved by an advanced algorithm 67 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Advanced HSDPA scheduling • An improved scheduling algorithm (Proportional Fair, PF) offers significant gain over a conventional algorithm (Round Robin, RR) • PF has a very good pricequality ratio – User equipment needs no changes – Node B’s need only minor changes 68 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Femtocells • More and more consumers want to use their mobile devices at home, even when there’s a fixed line available – Providing full or even adequate mobile residential coverage is a significant challenge for operators – Mobile operators need to seize residential minutes from fixed line providers, and compete with fixed and emerging Vo. IP and Wi. Fi services => There is trend in discussing very small indoor, home and campus Node. B layouts • Femtocells are cellular access points (for limited access group) that connect to a mobile operator’s network using residential DSL or cable broadband connections • Femtocells enable capacity equivalent to a full 3 G network sector at very low transmit powers, dramatically increasing battery life of existing phones, without needing to introduce Wi. Fi enabled handsets 69 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Femtocells • The study considers the system performance of an HSDPA network consisting of macro cells and very low transmit power (femto) cells • The impact of using 64 QAM in addition to QPSK and 16 QAM in order to benefit from the high SINR is studied • The network performance is investigated with different portions of users created in the buildings (0100%) 70 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008

Femtocells • Femtocells provide maximum of 1517 % gain to network throughput already without dedicated indoor users Table: Network throughput gain of femto cells to macro users Offered load Scheme • The gain is visible with high load in the network and comes directly from the increased number of access points in the network • Average load of a cell is decreased and users can be scheduled more often 71 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008 Medium High Congested Rake 1 x 1 3% 8% 15 % Rake 1 x 2 -1 % 19 % 13 % Equ 1 x 1 -2 % 18 % 15 % Equ 1 x 2 -1 % 3% 17 %

Femtocells • When the amount of dedicated indoor users increase, the gain of femto cells explodes • Gain is in the range of hundreds of percents even with small portion of indoor users 72 © Timo Nihtilä TLT-5606 Spread Spectrum Techniques / 25. 4. 2008