2015 SoftBank Trial Akihabara,Tokyo

Transcription

1 2015 SoftBank Trial Akihabara,Tokyo Adding street pole mounted Small Cells as a 2 nd LTE layer for the Macro deployment in a dense urban area Akihabara Tokyo 500mm Height limit Detached SBA 1

2 Trial Goals Evaluate the improvement in outdoor and indoor coverage and capacity gained by small cells in a hyper dense urban environment covered by Macro cells. Compare between the two frequency allocation scheme for heterogeneous networks: Dedicated channel Shared channel in accordance with interference coordination Evaluate different interference mitigation techniques: CoMP (between small cells) eicic (between Macro and small cells) Demonstrate V-RAN architecture based on L1 to L2/L3 split (the so called efapi interface) 2

3 Trial Area Akihabara Area: 400 x 200 m Products: 2 Macro-Cells Rooftop mounted (Band 42) 6 Small-Cells AirSynergy (Band 42) Street pole mounted SBA Antenna Antenna Direction Omni Direction X m 200m 3

4 Executed Tests - Outdoor Measure coverage and throughput compared to Macros only in the following scenarios: Drive Route Load: 0%, 100% HetNet Configuration: Dedicated Channel Both Macro and Small Cells utilizing different channels Co-Channel Macro and Small Cells utilizing the same channel Interference Coordination: CoMP between Small Cells: off, on eicic between Small Cells and Macro: Interpulated Fixed Locations Test Process: Drive test Fixed locations with multiple UEs 4

5 Executed Tests Indoor Measure indoor penetration and throughput in the following scenarios: Indoor3Site3Map Load: 0%, 100% SC3 HetNet Configuration: Co-Channel Macro and Small Cells utilizing the same channel Test Process: 4 fixed locations from outdoor to deep indoor 3m N 4 3m 8m 3 3m 2 50m 1 Ent. One test to compare to Macros only and CoMP

6 Outdoor Results Throughput No Load Throughput in main inner streets: Improves with small cells addition Directional has more optimized throughput Macro Only Macro + Small Cells Omni Macro + Small Cells Directional Main Inner Streets 6

7 Outdoor Coverage No Load Small Cells improves Throughput over Macro only Directional is better than Omni Outdoor Coverage No Load CDF 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Main Inner Streets Throughput CDF All SCs Omni With MCs All SCs Directional With MCs Macro Only Throughput [Mbps] Main Streets Avg. Throughput Macro Only Macro + Small Cells Omni Macro + Small Cells Directional Main Streets 7

8 Small Cells Indoor Coverage INDOOR PERFORMANCE (GACHA) 3 4 Counter Ent. 6m 2 10m 1 Ent. 8m SC4 N THROUGHPUT (MBPS) Outdoor by the window Indoor by the window Indoor middle of the room Macro Only Mc & SC's Indoor back of the room INDOOR PERFORMANCE (EURO CAFE) Macro Only THROUGHPUT (MBPS) Outdoor by the window Indoor by the window Indoor middle of the room Mc & SC's 4. Indoor back of the room 8

9 Shared vs. Dedicated Channel Allocation Inter Cell Interference is the main reason for performance degradation which limits the benefit of network densification. The results show cell edge gain of 90% when using dedicated channel and aggregate rate loss of over 20% Mitigating the interference between small cells and macro cells by using a dedicated channel is not recommended due to lower spectral efficiency CAPACITY DEDICATED VS. SHARED CHANNEL Aggregate Capacity (Mbps) Cell Edge Rate (Kbps) Dedicated Channel Shared Channel 9

10 Coordinated Multi Point Transmission (CoMP) Cooperative multi point transmission (CoMP) was introduced in 3GPP release 11 as an LTE-Advanced solution to improve cell edge throughput performance and consequently improve overall average cell throughput CoMP is based on cooperative transmission of data from multiple cells. Different schemes are supported, e.g. joint transmission, coordinated scheduling, dynamic point selection etc. Two general cases can be considered for CoMP UE is not aware of multi site transmission (Lite CoMP) TP1 PDSCH PDCCH PDSCH PDCCH TP2 TPs have same configuration of resource element mapping Applicable to JT and DPS May work with pre 3GPP release 11 UE UE is aware of multi point transmission (full 3GPP R11 CoMP) TPs have different configuration of resource element mapping Applicable to DPS Requires new DCI format and new TM TP1 PDSCH PDSCH PDCCH TP2 Requires new DL RS in order to provide CSI feedback for each TP independently 10

11 CoMP Clusters Structure SBA Antenna Antenna Direction Two CoMP clusters were created: PCI 11: SCs 3, 4, 7 Omni Direction 3 PCI 11 PCI 21: SCs 5, 6, 8 CoMP tests were executed on both CoMP clusters 2 PCI X 7 400m 8 200m 11

12 Throughput Results Co-Channel with 100% Load Macro has many dead spots in the main inner streets Small Cells improves the coverage and capacity CoMP improves and stabilizes the performance Macro Only Macro + Small Cells CoMP Off Macro + Small Cells CoMP On Main Inner Streets 12

13 CoMP Results Comparison Main Inner Streets Drive tests with one UE Macro has 26% no coverage Dedicated channel shows better results but uses 40MHz CoMP improves the results by 380% for Co-Channel 100% Drive Test Data (100% load) CDF 90% 80% 70% 60% 50% 40% 30% 20% 10% Dedicated Channel 100% Load CoMP ON Co Channel 100% Load CoMP ON Dedicated Channel 100% Load CoMP OFF Co Channel 100% Load CoMP OFF Macro Only 100% Load 0% Throughput [Mbps] 13

14 CoMP Results CoMP analysis show the following gains: Cell edge gain:140% Aggregate gain of 8% COMP PERFORMANCE (SHARED CHANNEL W/O EICIC) CAPACITY Aggregate Capacity (Mbps) No CoMP 110 Cell Edge Rate (Kbps) (Mbps) 3GPP CoMP 14

15 Indoor CoMP Results 100% Load Indoor CoMP shows 600% gain compared to Macro only and 50% - 100% compared to Small Cells with no CoMP Indoor Throughput Gain [Mbps] 3 1 Macros Only Macros + Small Cells CoMP ON x

16 Enhanced Inter Cell Interference Coordination (eicic) ICIC allows neighboring enodeb to coordinate their use of air interface resources to help avoid intercell interference. ICIC creates a tradeoff between improving the signal to noise ratio and reducing the quantity of resources available for transmission. ICIC can be applied in both the time and frequency domain. Time domain ICIC involves prioritizing the use of individual subframes within specific cells Frequency domain ICIC involves prioritizing the use of individual resource blocks within specific cells. The concept of ICIC was introduced in 3GPP release 8 Enhanced ICIC (eicic) was introduced in 3GPP release 10 3GPP release 10 introduces the concept of Almost Blank Subframes (ABS), which represents a form of time domain ICIC. The use of ABS involves enodeb reducing transmission during certain subframes ABS can be generated from normal subframes or from MBSFN subframes. Transmission is reduced by not scheduling PDSCH nor PMCH transmissions in those subframes. ABS using normal subframes does not eliminate interference caused by cell specific reference signals, synchronization signals and PBCH. Therefore, ABS using MBSFN subframes should be preferred. 16

17 eicic Results eicic performance was evaluated Cell edge gain: 45% Aggregate capacity gain: 15% EICIC PERFORMANCE CAPACITY Aggregate Capacity (Mbps) Cell Edge Rate (Kbps) ABS Ratio 0% ABS Ratio 40% 17

18 Combining CoMP and eicic CoMP may be combined with eicic (Time domain) ABS subframes exclusively allocated for small cells Macro (f1) Pico (f1) Non ABS ABS Non ABS ABS Non ABS Subframes For UE served by macro cell For CCU served by small cells (small cells may use CoMP for specific users if needed) ABS Subframes For CEU served by small cells Small cells may use CoMP for specific users if needed 18

19 Combined CoMP and eicic Substantial performance gain were observed when combining CoMP with eicic Cell edge gain: 260% Aggregate capacity gain: 20% COMP & EICIC COMBINED CAPACITY Aggregate Capacity (Mbps) No ICIC Cell Edge Rate (Kbps) CoMP & eicic 19

20 CoMP and eicic Gain Tested on all the full area see tested points 400% higher capacity compared to Macro only by using CoMP and eicic Network Aggregated Capacity (Mbps) Macro Only Dedicated Channel Co-Channel Co-Channel w.eicic No CoMP (3GPP CoMP) Macro Only 20

21 Conclusions The value of V-RAN Small Cells Rooftop macro cell layer is not able to provide an adequate coverage for street level and indoor locations in dense urban scenarios. Heterogeneous Networks is the solution An underlying layer of Small Cells can complete the required coverage in street level and indoor locations in dense urban scenarios: Small Cell can be used for coverage infill purposes in areas where macro cell layer signal is too low Small Cells densification increase network capacity, especially in hot spot locations. Small Cell gain compared to Macro only presented almost 400% outdoor gain and ~400% indoor gain Heterogeneous Networks Preferred Approach: Dedicated Channel is not an effective interference mitigation approach in terms of spectral efficiency Shared Channel approach should be preferred. It appears that the wider spectrum availability of the shared channel approach outperforms its higher interference level 21

22 Conclusions The value of V-RAN Small Cells Intra Small Cell Layer Interference Mitigation - CoMP CoMP proved to be a very effective solution for interference mitigation between Small Cells CoMP Aggregate Capacity Gain in Outdoor Locations observed: 400% compared to Macro 8% compared to Small Cells without CoMP This is aligned with the work of 3GPP (TR ), which indicated a possible gain of 2% to 10%. Outdoor Cell Edge CoMP Gain observed: 250% compared to Small Cells with no CoMP CoMP Indoor Gain observed: 600% compared to Macro 50% - 100% compared to Small Cells without CoMP Centralized CoMP: Centrailized CoMP based on pooling L2 processing and distributed L1 processing proved to be as efficient as pooling L1 and L2 processing using CPRI fronthaul interface The IP-CoMP architecture achieves similar CoMP gains with fronthaul rate which is an order of magnitude less than that of a CPRI based solution Inter Small Cell Macro Interference Mitigation - eicic Time Domain eicic is an effective way to mitigate interference between small cell layer and macro cell layer. As ABS solution is based on standard interface (X2), it is an attractive solution even when macro and small cells are provided by different vendors. The results indicated a potential gain of eicic in both average capacity (15%) and cell edge performance (45%) 22