Soft Handover Performance study in the Direct Sequence WCDMA Radio network Simulator

Transcription

1 Soft Handover Performance study in the Direct Sequence WCDMA Radio network Simulator Author: Yang Fang / Communication Laboratory Supervisor: Professor Sven-Gustav Häggman Instructor: PH.D Michael Hall 1

2 Table of Content Handover Introduction NETSIM Introduction NETSIM Structure and modules Introduction Channel Simulator Network Simulator Soft Handover algorithms Introduction Window-average algorithm Real-time algorithm Simulation result and Performance study Drop call trace tool Future work 2

3 Handover Introduction 3

4 Handover Introduction 4

5 Handover Introduction --- Interfrequency Handover 5

6 Handover Introduction --- Intrafrequency Handover 6

7 Handover Introduction --- Intersystem Handover 7

8 Handover Introduction --- Soft Handover 8

9 Handover Introduction --- Softer Handover 9

10 NETSIM Introduction 10

11 NETSIM Introduction NETSIM - simulation tool for study of planning methods and control algorithm for WCDMA cellular Radio Network Platform: Unix Language: C NETSIM can simulate: Voice and data service, packet switched traffic, circuit switched traffic, different user behavior, Radio network control functions (HO, Admission, Power Control) Simulation result: System capacity as a function of traffic, performance of network control algorithm, etc. 11

12 NETSIM Structure and Modules --- Structure Traffic Model Network Model Network Simulator WCDMA Simulation 1. Uplink algorithm; 2. Downlink algorithm; 3. Access Control 4. Admission control; 5. Soft handover 6. Power control Channel Simulator Network Performance Analysis NETSIM 12

13 Channel Simulator of NETSIM Current version using Raytracing model Impulse response is h( t) = K 1 k = 0 a δ ( t t k k ) exp( jθ ) Advantage: Model considered as accurate k Disadvantages: Computing intensive (large Memory and long simulation time required) Map of simulation environment 13

14 Network Simulator of NETSIM Traffic Model Network Model WCDMA Simulation Network Simulator 1. Uplink algorithm; 2. Downlink algorithm; 3. Access Control 4. Admission control; 5. Soft handover 6. Power control Network Performance Analysis 14

15 NETSIM Structure and Modules --- Modules Initialization Module Generation Module Mobile Station Traffic Module Mobility Module Propagation Module Radio Channel Interference Module Average Module Access Module Admission Module Active set Module Base Station Quality Module Power Module 15

16 Soft Handover algorithms Introduction 16

17 Handover Introduction --- Soft Handover 17

18 Measurement Quantity Soft Handover algorithm Introduction CPICH Window-average algorithm T T T As_Th + As_Th_Hyst AS_Th AS_Th_Hyst As_Rep_Hyst CPICH 2 CPICH 3 Time Cell 1 Connected Event 1A Add Cell 2 Event 1C Replace Cell 1 with Cell 3 Event 1B Remove Cell 3 Figure of window-average algorithm from 3GPP _3/02 18

19 Soft Handover algorithm Introduction --- Window-average algorithm flow chart Begin M eas_sign > Best_Ss As_Th as_th_hyst for a period of T Yes No (Event 1B) Remove Worst_Bs in the Active Set M eas_sign > Best_Ss As_Th + as_th_hyst for a period of T No Yes (Event 1A) Active Set Full Yes No Best_Cand_Ss > W orst_old_ss + As_Rep_Hyst for a period of T No Yes (Event 1C) Add Best BS in Active Set and Remove W orst B s from th Active Set Add B est_bs in the Active Set Flow chart of window-average algorithm from 3GPP _3/02 19

20 Parameters for Window-average algorithm Parameters: AS_Th AS_Th_hyst AS_Rep_hyst HO_Add_time HO_Drop_time Threshold of Marco-diversity gain in Window-average algorithm Hysteresis of AS_Th Replacing Hysteresis in Window-average algorithm Evaluating window size to add candidate to active set list Evaluating window size to drop one from active set list 20

21 Soft Handover algorithm Introduction --- Real-time algorithm Always connect to the cells with better or best signal quality Swap the cells in the active set frequently Response quickly to the change of the communication channel No window to evaluate the receiving signal Soft handover gain is fixed (equal to the Marco-diversity gain) 21

22 Parameters for Real-time algorithm Parameters: AS_3_ratio AS_2_ratio Marco-diversity gain when using 3 active set in the Soft handover procedure in Real-time algorithm Marco-diversity gain when using 2 active set in the Soft handover procedure in Real-time algorithm 22

23 Simulation result and Performance study 23

24 Performance study for different algorithms --- Window-average algorithm (1-1) AS active Threshold 3.98(equal to 6 db) AS active Threshold Hysteresis 1.58(equal to 2 db) AS active Replacement Threshold Hysteresis 3.98(equal to 6 db) AS active Handover add window size 0.5(second) AS active Handover drop window size 0.5(second) Group 1 Parameters set 24

25 Performance study for different algorithms --- Window-average algorithm (1-2) Successful call vs. drop call (Window size = 0.5 s, AS_Th = 6 db, AS_Rep_Th = 6dB) Percentage of acting different radio links in the SHO procedure (Window size = 0.5 s, AS_TH = 6 db, As_Rep_Th = 6 db) % % 98.23% 99.17% 90.00% 80.00% 79.50% 79.19% 90.00% 80.00% Percentage (%) 70.00% 60.00% 50.00% 40.00% 30.00% 20.00% 20.50% 20.81% Calls ended normally Nr. of quality fail calls Percentage(%) 70.00% 60.00% 50.00% 40.00% 30.00% 20.00% Acting 3 radio link case in this simulation loop Acting 2 radio link case in this simulation loop Acting 1 radio link case in this simulation loop 10.00% 0.00% AS=3 AS= % 0.00% 0.03% 1.75% 0.83% AS=3 AS=2 Maximum active set size Maximum active set size Simulation result of parameter set group1 25

26 Performance study for different algorithms --- Window-average algorithm (2-1) AS active Threshold 2.512(equal to 4 db) AS active Threshold Hysteresis 1.58(equal to 2 db) AS active Replacement Threshold Hysteresis 2.512(equal to 4 db) AS active Handover add window size 0.5(second) AS active Handover drop window size 0.5(second) Group 2 Parameters set 26

27 Performance study for different algorithms --- Window-average algorithm (2-2) Successful call vs. Drop call (Window size = 0.5 s, AS_Th = 4 db, AS_Rep_Th = 4 db ) Percentage of acting different radio links case in SHO procedure (Window size = 0.5 s, AS_Th = 4 db,as_rep_th = 4 db) % % 99.31% 98.97% 90.00% 80.00% 79.22% 81.65% 90.00% 80.00% Percentage (%) 70.00% 60.00% 50.00% 40.00% 30.00% 20.00% 20.78% 18.35% Calls ended normally Nr. quality failures ul Percentage (%) 70.00% 60.00% 50.00% 40.00% 30.00% 20.00% Acting 3 radio link case in this simulation loop Acting 2 radio link case in this simulation loop Acting 1 radio link case in this simulation loop 10.00% 0.00% AS=3 AS=2 Maximum Active set size 10.00% 0.00% 0.01%0.69% 1.03% AS=3 AS=2 Maximum Active Set size Simulation result of parameter set group2 27

28 Performance study for different algorithms --- Window-average algorithm (3-1) AS active Threshold 2.512(equal to 4 db) AS active Threshold Hysteresis 1.58(equal to 2 db) AS active Replacement Threshold Hysteresis 2.512(equal to 4 db) AS active Handover add window size 0.1(second) AS active Handover drop window size 0.1(second) Group 3 Parameters set 28

29 Performance study for different algorithms --- Window-average algorithm (3-2) Successful calls vs. Drop calls (Window size = 0.1 s AS_Th = 4dB, AS_Rep_Th= 4 db) Percentage of acting different radio links case (Window size = 0.1 s AS_Th = 4dB, AS_Rep_Th= 4 db) 90.00% 80.00% 70.00% 72.83% 80.95% % % 97.97% 96.82% Percentage (%) 60.00% 50.00% 40.00% 30.00% 20.00% 10.00% 0.00% 27.17% AS= % AS=2 Calls ended normally Nr. quality failures ul Percentage (%) 80.00% 60.00% 40.00% 20.00% 0.00% 0.07% 1.96% 3.18% AS=3 AS=2 Acting 3 radio link case in this simulation loop Acting 2 radio link case in this simulation loop Acting 1 radio link case in this simulation loop Maximum active set size Maximum active set size Simulation result of parameter set group3 29

30 Performance study for different algorithms --- Window-average algorithm summary(1) Comparison table of successful calls and drop calls (MaximumActive set size is 3) Percentage of acting different radio links case with different parameters sets (Maximumactive set size is 3) Percentage 90.00% 80.00% 70.00% 60.00% 50.00% 40.00% 30.00% 20.00% 10.00% 0.00% 79.50% 79.22% 20.50% 20.78% WS = 0.5; As_Th= 6db WS = 0.5; As_Th= 4db 72.83% 27.17% WS = 0.1; As_Th= 4db Calls ended normally Nr. of quality failures Percentage % % 80.00% 60.00% 40.00% 20.00% 0.00% 98.23% 99.31% 97.97% 0.03% 1.75% 0.01% 0.69% 0.07% 1.96% WS = 0.5; As_Th= 6db WS = 0.5; As_Th= 4db WS = 0.1; As_Th= 4db Acting 3 radio link case in this simulation loop Acting 2 radio link case in this simulation loop Acting 1 radio link case in this simulation loop Different Parameters sets Different parameters sets Simulation result of different parameter sets of Window-average algorithm (AS=3) 30

31 Performance study for different algorithms --- Window-average algorithm summary(2) Comparison Table of successful call and drop call (Maximumactive set size is 2) Percentage of acting different radio links case with different parameters sets (Maximumactive set size is 2) 90.00% 80.00% 70.00% 79.19% 81.65% 80.95% % % 99.17% 98.97% 96.82% Percenntage (%) 60.00% 50.00% 40.00% 30.00% 20.00% 10.00% 0.00% 20.81% WS = 0.5; As_Th= 6db 18.35% 19.05% WS = 0.5; As_Th= 4db WS = 0.1; As_Th= 4db Calls ended normally Nr. of quality failures Percentage (%) 80.00% 60.00% 40.00% 20.00% 0.00% 0.83% 1.03% 3.18% WS = 0.5; As_Th= 6db WS = 0.5; As_Th= 4db WS = 0.1; As_Th= 4db Acting 2 radio link case in this simulation loop Acting 1 radio link case in this simulation loop Different parameters sets Differnt parameters sets Simulation result of different parameter sets of Window-average algorithm (AS=2) 31

32 Performance study for different algorithms --- Window-average algorithm summary(3) Conclusion: The performance is a little better when the active set size is 2 In most time of the call procedure, the communication between the MS and BS only using one radio link It s difficult for find the optimal parameters set for the Window-average algorithm, we have to try a lot of parameters sets to get one better solution for the Window-average algorithm 32

33 Performance study for different algorithms --- Real-time algorithm simulation result Successful call rate vs. drop call rate (Real-time algorithm) Percentage of acting radio links in the simulation loop (Real-time algorithm) % 90.00% 99.11% % % 90.00% 87.29% 88.24% 80.00% 80.00% percentage(%) 70.00% 60.00% 50.00% 40.00% 30.00% Calls ended normally Nr. of quality fail calls Percentage(%) 70.00% 60.00% 50.00% 40.00% 30.00% Acting 3 radio link case in this simulation loop Acting 2 radio link case in this simulation loop Acting 1 radio link case in this simulation loop 20.00% 10.00% 0.00% 0.89% 0.00% AS=3 AS=2 Maximum Active set size 20.00% 10.00% 0.00% 4.09% 8.62% 11.76% AS=3 AS=2 Maximum active set size Simulation result of Real-time algorithm 33

34 Performance study for different algorithms --- comparison between two algorithms Conclusion: The Real-time algorithm is better than Window-average algorithm The Real-time algorithm always adopt the better channel for the conversation. But the window-average algorithm need some average window to adopt the optimal link 34

35 Drop call trace tool 35

36 Drop call Trace tool --- Network drop call situation in Time 510 Time:510 Relatioinship between Base Station SIR and Drop call Rx value (db) Value Drop call Nr The 0 Base station The 1 Base station The 2 Base station The 3 Base station Base station The 4 Base station The 5 Base station The 6 Base station The drop call in SHO with 3AS The drop call in SHO with 2AS The drop call in SHO with 1AS The Base station's average SIR 36

37 Drop call Trace tool --- Network drop call situation in Time 520 Time:520 Relationship between Base Station and Drop call Rx value (db) Value Drop call Nr The 0 Base station The 1 Base station The 2 Base station The 3 Base station Base station The 4 Base station The 5 Base station The 6 Base station The drop call in SHO with 3AS The drop call in SHO with 2AS The drop call in SHO with 1AS The Base station's average SIR 37

38 Future Work 38

39 Future work More simulations with different parameters set are needed The Packet data service performance with different SHO algorithms is need to be investigated More drop call trace analysis tool need to be integrated in NETSIM 39

40 Thank you! Questions? 40