UTRAN Radio Resource Management

Transcription

1 UTRAN Radio Resorce Management BTS 3 BTS 1 UE BTS 2 Introdction Handover Control Soft/Softer Handover Inter Freqency Handover Power Control Closed Loop Power Control Open Loop Power Control Interference Management Load Control Call Admission Control Congestion Control Packet Data Transmission Packet Data Control Dynamic Schedling

2 References H. Holma, A. Toskala (Ed.), WCDMA for UMTS, Wiley, 5th edition, Wiley, 2010 H. Kaaranen, et.al., UMTS Networks: Architectre, Mobility and Services, Wiley, (see chapter 4) A. Viterbi: CDMA: Principles of Spread Spectrm Commnications, Addison Wesley, J. Laiho, A. Wacker, T. Novosad (ed.): Radio Network Planning and Optimisation for UMTS, Wiley, 2001 T. Ojanperä, R. Prasad, Wideband CDMA for Third Generation Mobile Commnication, Artech Hose, R. Prasad, W. Mohr, W. Konhäser, Third Generation Mobile Commnications Systems, Artech Hose, March GPP standards: TS : Physical Layer Procedres (esp. power control) TR : Radio Resorce Management Strategies TR : RF System Scenarios UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

3 RRM High-Level Reqirements Efficient se of limited radio resorces (spectrm, power, code space) Minimizing interference Flexibility regarding services (Qality of Service, ser behavior) Simple algorithms reqiring small signalling overhead only Stability and overload protection Self adaptive in varying environments Allow interoperability in mlti-vendor environments Radio Resorce Management algorithms control the efficient se of resorces with respect to interdependent objectives: cell coverage cell capacity qality of service UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

4 RRM Components Core Network/ other RNCs Radio Resorce Management Handover Control Load Control Packet Data Control typically in RNC Power Control Medim Access Control typically in NodeB Physical layer UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

5 Handover Control: Basics General: mechanism of changing a cell or base station dring a call or session Handover in UMTS: UE may have active radio links to more than one Node B Mobile-assisted & network-based handover in UMTS: UE reports measrements to UTRAN if reporting criteria (which are set by the UTRAN) are met UTRAN then decides to dynamically add or delete radio links depending on the measrement reslts Types of Handover: Soft/Softer Handover (dedicated channels) Hard Handover (shared channels) Inter Freqency (Hard) Handover Inter System Handover (e.g. UMTS-GSM) Cell selection/re-selection (inactive or idle) All handover types reqire heavy spport from the UMTS network infrastrctre! UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

6 Macro Diversity & Soft Handover (Wrap-Up) NodeB 1 NodeB 2 UE Downlink: combining in the mobile station Uplink: combining in the base station and/or radio network controller UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

7 Soft/ Softer Handover In soft/softer handover the UE maintains active radio links to more than one Node B Combination of the signals from mltiple active radio links is necessary Soft Handover The mobile is connected to (at least) two cells belonging to different NodeBs In plink, the signals are combined in the RNC, e.g. by means of selection combining sing CRC Softer Handover The mobile is connected to two sectors within one NodeB More efficient combining in the plink is possible like maximm ratio combining (MRC) in the NodeB instead of RNC Note: In plink no additional signal is transmitted, while in downlink each new link cases interference to other sers, therefore: Uplink: HO general increase performance Downlink: Trade-off UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

8 Soft and Softer Handover in Practice UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

9 Soft Handover Control soft handover area NodeB 1 UE NodeB 2 Measrement Qantity CPICH 1 CPICH 2 d add d drop T link Measrement qantity, e.g. E C /I 0 on CPICH Relative thresholds d add & d drop for adding & dropping Preservation time T link to avoid ping-pong effects Event triggered measrement reporting to decrease signalling load Link to 1 Link to 1 & 2 Link to 2 time UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

10 Soft Handover Simlation Reslts 25% Otage Probability (Blocking and Dropping) 20% 15% 10% 5% 1 link max 2 SHO links max 4 SHO links max 6 SHO links 0% Offered Traffic [Erlang per site] Soft handover significantly improves the performance, bt UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

11 Soft Handover Simlation Reslts II 2 Mean Nmber of Active Links 1,5 1 0, Max. Active Set Size the overhead de to simltaneos connections becomes higher! UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

12 Inter-Freqency Handover Hierarchical cell strctre (HCS) Hot-spot Macro Micro Macro f 1 f 2 f 1 Hot spot f 1 f 2 f 1 f 1 Handover f 1 Û f 2 always needed between layers Hard handover Handover f 1 Û f 2 needed sometimes at hot spot Inter-freqency measrements of target cell needed in both scenarios Mobile-assisted handover (MAHO) slotted (compressed) mode for inter-freqency measrements to find sitable target cell also spports GSM system measrements Database assisted handover (DAHO) no measrements performed on other freqencies or systems se cell mapping information stored in data base to identify the target cell UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

13 Cell Selection/Re-selection Handover when UE is idle or there is no active connection between UE and UTRAN Goal: find a sitable cell to camp on Cell selection border The cell to camp on is chosen by the UE on measred link qality Q, e.g. E C /I 0 on CPICH (after cell-search) Cell re-selection with hysteresis H to avoid ping-pong effects NodeB 1 Q 1 Q 2 UE Q 1 > Q min Q 1 > Q 2 NodeB 2 Additional offsets for Q on different freqencies, e.g. to spport hierarchical cells Mapping fnctions for Q between UMTS and GSM to spport priorities Cell selection and re-selection mainly performed internally in UE, bt controlled by UTRAN with broadcast of neighbor cell freqencies and control parameters (hysteresis, mapping, etc.) UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

14 Power Control: Basics Controls the setting of the transmit power in order to: Keep the QoS within the reqired limits, e.g. data rate, delay and BER Minimise interference, i.e. the overall power consmption Power control handles: Path Loss (Near-Far-Problem), Shadowing (Log-Normal-Fading) and Fast Fading (Rayleigh-, Ricean-Fading) Environment (delay spread, UE speed, ) which implies different performance of the de-interleaver and decoder Uplink: per mobile Downlink: per physical channel Three types of power control: Inner loop power control Oter loop power control (SIR-target adjstment) Open loop power control (power allocation) Downlink power overload control to protect amplifier Gain Clipping (GC) Aggregated Overload Control (AOC) UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

15 Near-Far Problem Power Control (Wrap-Up) UE 1 Near-Far Problem: Spreading seqences are not orthogonal (mlti-ser interference) Near mobile dominate Signal to interference ratio is lower for far mobiles and performance degrades NodeB The problem can be resolved throgh dynamic power control to eqalize all received power levels AND/OR UE 2 By means of joint mlti-ser detection UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

16 Closed Loop Power Control Closed loop power control is sed on channels, which are established in both directions, sch as DCH There are two parts Inner Loop Power Control (ILPC): receiver generates p/ down commands to incrementally adjst the senders transmit power Oter Loop Power Control (OLPC): readjsts the target settings of the ILPC to cope with different fading performance SIR > SIR target? target adjstment BLER target UE Inner Loop (1500 Hz) Oter Loop ( 100Hz) RNC control command: Up/Down NodeB Example: Uplink Closed Loop Power Control UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

17 Impact of Power Control 8 speed = 3 km/h 7 E b /N 0 [db] power/ fading [db] time [sec] Example: UMTS Closed Loop Power Control in the slow fading channel UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

18 Power Control Performance Reqired UL SIR [db] PedA VehA Velocity [km/h] SIR reqirement strongly depends on the environment (de to different fast fading conditions Jakes models) Þ oter loop power control needed to adapt SIR UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

19 Open Loop Power Control Open loop power control is sed on channels that cannot apply closed loop power control, e.g. RACH, FACH The transmitter power is determined on the basis of a path loss estimate from the received power measre of the opposite direction To avoid excessive interference, probes with incremental power steps ntil a response is obtained: power ramping UE NodeB Open Loop Power Control on RACH UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

20 CDMA Overload CDMA systems tend to become nstable More traffic increases the interference More interference reqires higher power More power increases the interference Methods are reqired to limit the system load Restrict the access to the system Overcome overload sitations UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

21 Interference in CDMA Networks Interference Inter-Symbol Interference (ISI) Problem Delayed components from the same ser signal interfere de to mltipath propagation Mltiple Acces Interference MAI Different ser signals interfere dependent on the access scheme Intra-Cell Interference Inter-Cell Interference Interference cased by the sers belonging to same cell Interference cased by the sers belonging to neighbor cells. Freqency rese factor is one CDMA is sbject to high mltiple access interference Soft capacity: CDMA capacity (e.g. nmber of sers) determined by the interference is soft Handling of interference is the main challenge in designing CDMA networks UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

22 Cell Breathing CDMA systems: cell size depends on the actal loading Additional traffic will case more interference If the interference becomes too strong, sers at the cell edge can no more commnicate with the basestation CDMA interference management Restriction of the sers access necessary Cell breathing makes network planning difficlt Example: cell brething with increasing traffic UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

23 Cell Breathing (contd.) Coverage depending on load: load cases interference, which redces the area where a SIR sfficient for commnication can be provided coverage low load coverage medim load coverage high load shadowed area: connection maybe lost UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

24 Coverage vs. Capacity Capacity depends on: QoS of the sers (data rate, error performance (bit-error-rate)) User behavior (activity) Interference (ot of cell) Nmber of carriers/ sectors Coverage (service area) depends on: Interference (intra- & inter-cell) + noise Pathloss (propagation conditions) QoS of the sers (data rate, error performance (bit-error-rate)) Ths, trade-off between capacity and coverage UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

25 Coverage vs. Capacity kbps circit switched service capacity verss maximm cell radis 3 Maximm cell radis (km) Downlink 0.5 Uplink Erlangs (2% GOS) Downlink limits capacity while plink limits coverage Downlink depends more on the load (sers share total transmit BS power) UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

26 Example of Coverage and Best Server Map coverage map best server map Application: RF engineering (cell layot) Legend: violet indicates high signal level, yellow indicates low level Application: HO decision Legend: color indicates cell with best CPICH in area UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

27 Load Control: Basics Main objective: Avoid overload sitations by controlling system load Monitor and controls radio resorces of sers Call Admission Control (CAC) Admit or deny new sers, new radio access bearers or new radio links Avoid overload sitations, e.g. by means of blocking the reqest Decisions are based on interference and resorce measrements Congestion Control (ConC) Monitor, detect and handle overload sitations with the already connected sers Bring the system back to a stable state, e.g. by means of dropping an existing call UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

28 Resorce Consmption Service/BLER-dependent resorce consmption Uplink example: Service I: Voice R b = 12.2kbps, E b /N t = 5dB a I = 0.99% Service II: Data R b = 144kbps, E b /N t = 3.1dB a II = 7.11% a In downlink there is additional dependency on the location of the ser Cell center low consmption Cell edge high consmption UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

29 Admission/ Congestion Control Basic algorithm Admission control is triggered when load ³ thr_cac New sers are blocked Existing sers are not affected as long as load < thr_conc Congestion Control is triggered when load ³ thr_conc Redce consmption of one or several sers Simple action: drop the ser Repeat ntil load < thr_conc UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

30 Call Admission Control: Simlation Reslts I Tradeoff between blocking and dropping Example: 64k per ser, rban 50% 45% 40% thr_cac = 50% thr_cac = 75% thr_cac = 90% 20% 18% 16% thr_cac = 50% thr_cac = 75% thr_cac = 90% Blocking Probability 35% 30% 25% 20% 15% Dropping Probability 14% 12% 10% 8% 6% 10% 4% 5% 2% 0% % Offered Traffic [Erlang per site] Offered Traffic [Erlang per site] UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

31 Call Admission Control: Simlation Reslts II Cell load depending on CAC threshold Example: 64k per ser, rban 90% 80% Cell Loading 70% 60% 50% 40% 30% 20% 10% 0% Offered Traffic [Erlang per site] thr_cac = 50% thr_cac = 75% thr_cac = 90% UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

32 Packet Data Control: Channel Switching Flexibility of packet services Asymmetrical data rates Very low to very high data rates Control information/ser information Efficient transmission making good se of CDMA characteristics Dedicated channel (DCH) Minimise transmission power by closed-loop power control Independence between plink and downlink capacity Common channel Random access in the plink (RACH) Dynamic schedling in the downlink (FACH) Adaptive channel sage depending on traffic characteristics Infreqent or short packets Þ Common channel (Cell_FACH) Freqent or large packets Þ Dedicated channel (Cell_DCH) No packet transmission Þ UE stand by mods (URA_PCH) UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

33 Channel Switching Example CELL_DCH CELL_FACH CELL_DCH DCH Active Time Page Download Time Reading Time Chatty Applications Example: Web service Chatty apps.: keep alive message, stock tickers, etc. (e.g. 100 bytes every 15 sec) Second stage: when no activity in CELL_FACH then switch to URA_PCH UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

34 Power Control vs. Rate Adaptation NodeB high data rate area UE 1 low data rate area UE 2 Power Control: Balances ser received qality (BLER, SIR) Users at cell center get less share of BTS transmit power assigned than at cell edge Occrrence of power overload Rate Adaptation: Transmit power ~ data rate Users at cell edge get lower data rate assigned than at cell center Redces also power overload On DCH combination of power control and rate adaptation Rate assignment at begin of a transmission based on load and ser location Rate adaptation when ongoing transmission according to power consmption and overload Based on RRC-signaling (time horizon: 100msec 10sec) UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

35 Rate Adaptation Performance UMTS_rban, 50 kbyte UMTS_rban, 50 kbyte 40% 35% 384k 64k adaptive % 6 384k Otage Probability 25% 20% 15% 10% Mean Delay [sec] k adaptive 5% 1 0% Cell Throghpt [kbit/sec] Cell Throghpt [kbit/sec] Rate adaptation significantly improves the RRM performance on DCH. UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct

36 Dynamic Schedling NodeB Flow #3 Sample Flow Flow #1 Flow #2 UE 2 UE 1 Statistical mltiplexing of data packets from different data flows on one shared medim, e.g. on DSCH or HSDPA Schedling with time-horizon of 2msec 1sec Optimised sage of radio resorces Exploitation of the short-term variations on the radio channels (opportnistic schedling) Can provide certain degree of QoS UE 3 UMTS Networks Andreas Mitschele-Thiel, Jens Meckenheim Oct