hash stringlengths 32 32 | doc_id stringlengths 5 12 | section stringlengths 5 1.47k | content stringlengths 0 6.67M |
|---|---|---|---|
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.7.8.4.2 TX architecture | Figure 75 shows the proposed TX architecture of the dual carrier mobile.
Figure 75: Architecture of uplink dual carrier mobile
Note that only a single band is shown in the above figure for the sake of simplicity. Additional switch positions at the antenna switchplexer could be used for inputs/outputs of other bands. Th... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.7.8.4.3 Throughput in coverage limited scenario | In this subclause the median uplink throughput (at received signal level of -98 dBm) as a function of the number of uplink time slots is compared for single carrier EGPRS and dual carrier EGPRS mobiles. For this purpose, the link level throughput curves shown in [8] are used.
It is assumed that the mobile station follo... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.7.9 Discussion on Uplink Coverage | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.7.9.1 Introduction | In this subclause coverage aspects for some uplink proposals are discussed. Coverage analysis includes multislot power reduction with different multislot profiles and also impact of insertion losses e.g. due to duplexers and isolators. Analyzed proposals include DSR, MDSR, UL DC with independent carriers, modified UL D... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.7.9.2 Assumptions for power reductions and power consumption | Multi slot power reduction was taken into account by applying multi slot profiles (0 and 3) for all the cases. For 8PSK modulation in the case of EGPRS and DSR 4 dB power reduction was applied related to GMSK [9]. For MDSR 16QAM 6 dB power reduction was applied at highest power level and 4 dB for other levels [10].
Th... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.7.9.3 Receiver and Network model | The BTS receiver and network model was as in [10]. Median RX level under interest was -98 dBm and RX level at cell border was -108 dBm.
Figure 79: Throughput versus received signal level, TU3iFH, NF=5dB |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.7.9.4 Results | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.7.9.4.1 Cell border | Figures 80 and 81 show throughput versus timeslots at cell border for multislot profile 0 and 3 respectively.
MDSR seems to outperform other schemes at cell border.
Type-2 MS with multislot profile 3 would need 6 or more slots to exceed throughput obtained by EGPRS already with 4 slots.
Dual carrier seems not to provid... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.7.9.4.2 Median coverage | Figures 82 and 83 show median uplink throughput versus number of time slots used for multislot profile 0 and 3 respectively.
Figure 82: Median throughput (-98 dBm)
with multislot profile 0
Figure 83: Median throughput (-98 dBm)
with multislot profile 3
Figures 84 and 85 show median uplink throughput gain related to EGP... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.7.10 Improvements for DTM and MBMS | This subclause highlights the additional gains for MBMS and DTM in terms of added flexibility for resource allocation and additional downlink throughput that can be obtained if mobile stations support dual carrier on the uplink. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.7.10.1 Assumptions | The following assumptions are made about the dual carrier technology applicable to the MS:
- The MS cannot transmit and receive in the same slot.
- The MS can receive on both carriers on a given timeslot. (Dual Carrier in the downlink).
- The MS cannot transmit on any carrier on the slot immediately following a slot it... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.7.10.2 Gains for DTM Multislot Capacity | With dual carrier only in the downlink, it is shown in [7] that the maximum capacity for a downlink biased DTM call is 10 TSs for reception and up to 2 TSs for transmission (with sum = 12). For an uplink biased allocation the corresponding figures are shown to be up to 4 TSs for reception and 5 TSs for transmission ( w... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.7.10.3 CS Connection setup while in packet transfer mode | According to the existing Rel-6 DTM behaviour, if the network wishes to establish an SDCCH for CS connection set-up before allocating a TCH, the likelihood is that the existing packet resources will first have to be moved to be next to an SDCCH, and then moved again when a TCH is required for speech. This involves two ... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.7.10.4 Gains for MBMS + CS | Increased capabilities of Mobile stations like dual carrier in DL/UL etc may in the future allow users to receive MBMS and CS calls simultaneously. Currently, it is not possible to support a CS call for any mobile which is already receiving the MBMS session with feedback if the mobile is not capable of transmitting or ... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.7.11 Performance enhancing features | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.7.11.1 Intercarrier Interleaving | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.7.11.1.1 Introduction | Diagonal intercarrier interleaving can be used to gain additional frequency diversity for dual carrier transmission on the uplink (see clause 10 on the use of intercarrier interleaving for reducing latency). This concept is likely not applicable as-it-is on the downlink because, downlink is a shared channel and old and... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.7.11.1.2 Link level gains by intercarrier interleaving | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.7.11.1.2.1 Used Simulation Parameters | Simulations are performed for TU50 channel model for receiver sensitivity case. Initial simulations show the performance comparison for MCS-5 and MCS-6 with and without intercarrier interleaving. Both ideal frequency hopping and no frequency hopping cases are simulated. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.7.11.1.2.2 Impairments | Transmitter and receiver impairments have not been included in the currently presented set of simulation results. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.7.11.1.2.3 Simulation Results | The results presented here are only for the sensitivity limited scenarios as the sensitivity limited scenarios are of main concern when applying dual carrier on the uplink as there is a reduced power transmission on the uplink.
It can be observed in figure 91 for a non frequency hopping channel and in figure 92 for a f... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.7.11.2 Advanced Coding Schemes | Advanced coding schemes like turbo coding [5] etc are known to perform better with longer input block lengths. Further link level gains are foreseen with dual carrier transmission on uplink with turbo coding. The idea is to compensate the loss in coverage due to reduced power transmission of the MS using additional lin... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.8 Impacts to the BSS | Multi-carrier is expected to have no impact on EDGE transceivers, but the BSS needs to perform data transfer (possibly including incremental redundancy transmission), resource allocation and link control for more than one carrier.
Dual Carrier in the UL enables maximum reuse of the existing BSS infrastructure, avoiding... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.9 Impacts to the Core Network | No changes are expected to the core network except that new capabilities shall be signalled by the MS to the network. For DC in the uplink, no further changes to Gb interface are required. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.10 Radio network planning aspects | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.10.1 Analysis for Option C | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.10.1.1 Introduction | Due to the issues and complexity regarding dual carrier in uplink with two separate transmitters, the use of a wideband transmitter has been proposed in order to simplify the implementation (see subclause 7.7.11.4).
The maximum carrier separation for a wideband transmitter is estimated to be 1 MHz. The impact of this l... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.10.1.2 Legacy Frequency Planning | There exists a number of different frequency planning techniques all depending on numerous parameters as e.g. geographical environment, traffic load, frequency hopping etc. In this report we will focus on two main techniques - Fractional Load Planning (FLP) and Multiple Reuse Planning (MRP).
FLP: A FLP network is plann... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.10.1.3 Impact of Wideband transmitter on legacy frequency planning | The maximum carrier separation in a wideband transmitter is e.g. 1 MHz, due to InterModulation (IM) products and the linearization of the PA. From a system point-of-view there is a requirement on the minimum carrier separation in order to ensure a certain quality level in a cell. The minimum carrier separation is 400 k... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.10.1.3.1 FLP-1/1 or 1/3 | When considering frequency planning for a wideband transmitter there is no difference between the 1/1-frequency reuse or 1/3. Two different cases have been considered for the FLP network: in the first case both carriers are placed in the same TCH hopping group and in the second case one of the carriers is placed on the... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.10.1.3.1.1 Both Carriers in the TCH hopping group | In figure 94 a blocked configuration is used and each cell has two transceivers in the TCH hopping group. There are two different Mobile Allocation alternatives, and only alternative 2 is applicable for dual carrier in the uplink with a wideband transmitter.
Figure 94: Different Mobile Allocation alternatives in a FLP ... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.10.1.3.1.2 One of the carriers on the BCCH | If the two carriers are separated between the BCCH and the TCH hopping group it will be more or less impossible to fulfil the requirement on the maximum carrier separation. If the requirement is to be fulfilled in this configuration the MS must be able to support a carrier separation that is equal to the total bandwidt... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.10.1.3.2 MRP | Most MRP networks are including the BCCH frequency in the hopping set, the separation between two carriers in a MRP cell is therefore >> 1 MHz. This makes it impossible to include dual carrier in uplink using a wideband transmitter in a MRP network.
However, if a re-planning of the frequencies is allowed it could be po... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.10.1.4 Extended frequency allocation | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.10.1.4.1 Introduction | A conclusion from subclause 7.10.1.3 is that in some networks it will be rather tricky to implement dual carrier in uplink using a wideband transmitter without any impact to the legacy frequency planning. A solution that removes this obstacle is therefore investigated in this subclause. Note that some operators do not ... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.10.1.4.2 Description of Extended Frequency Allocation | The idea with the solution is to temporarily assign an uplink frequency for carrier two that is within the maximum carrier separation to carrier one. This temporarily used frequency does not need to be a frequency that is normally allocated in the cell. This solution will work in all kinds of networks and there is no i... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.10.1.4.3 Impact of EFA on the BTS | EFA shall only be applied to the timeslots that are used by dual carrier in uplink and the duplex distance can therefore be different between the timeslots within a TRX. The impact of changed duplex distance can be implementation dependent and needs to be investigated by each network vendor. In the Ericsson BTS there i... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.10.1.4.2.1 Non-frequency hopping | The use of EFA with a fixed frequency has already been described in subclause 7.10.1.4.2, but in this subclause is an example with EFA in a network with a frequency reuse of 12 presented. Each cell in table 32 has two TRXs with a frequency reuse of 12, but when resources are assigned for dual carrier in uplink this reu... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.10.1.4.2.2 Frequency Hopping | It has already been shown in subclause 7.10.1.3 that dual carrier in uplink could work in a network using synthesized frequency hopping and legacy frequency allocation. However, it could be needed to decrease the number of frequencies allocated per TRX in order to avoid the "wrap-around" problem. If this reduction in a... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.10.1.4.3 Separation of carriers | Different separation between the carriers can be needed in order to avoid co-channel and adjacent channel interference within a site. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.10.1.5 Evaluation of network performance | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.10.1.5.1 Setup | Simulations have been performed to evaluate the system impact from the introduction of dual carrier in uplink with EFA. The following settings have been used.
Frequency reuse
12
Sectors (cells) per site
3
Number of TCH frequencies
36 (7.2 MHz), not including BCCH
Number of cells simulated
75 + wrap around
IRC
No
Cell r... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.10.1.5.2 Results | The results are presented in the form of a CDF of the uplink C/I distribution for the speech users of each case. C/I is averaged over measurement periods of 480 ms.
Figure 105: C/I distribution for speech users in a network with a frequency reuse of 12
The dashed line is for dual carrier in uplink with EFA
From figure ... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.11 Impacts to the specifications | The impacted 3GPP specifications are listed in 33.
Table 33: Impacted 3GPP specifications
Specification
Description
Comments
43.055
DTM Stage 2
43.064
GPRS Stage 2
45.001
Physical layer one radio path; general description
45.002
Multiplexing and multiple access on the radio path
45.005
Radio transmission and reception
... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 7.12 References | [1] Padhye & al. "Modeling TCP Reno Performance: A simple model and its empirical validation" IEEE/ACM Transaction on Networking, vol. 8, no. 2, April 2000.
[2] AHGEV-020, "Multi-carrier GERAN", 3GPP GERAN-EV Ad Hoc #1.
[3] GP-052723, "GERAN Evolution - Proposed Text on Dual Carrier in the Uplink for Technical Report",... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8 Higher Order Modulation and Turbo Codes | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.1 Introduction | Higher order modulations should be considered as a candidate to increase peak rates and, more importantly, to increase the mean bit rates. Convolutional coding is currently used for MCS coding for packet service over EGPRS. As a replacement to these within Future GERAN Evolution, Turbo Coding schemes can also be consid... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.2 Concept Description | The coding and modulation schemes that are already available for the current EDGE system are enhanced with the introduction of higher order modulations and Turbo coding. Since the higher order modulations enable higher data rates, new coding schemes may also be introduced. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.2.1 Higher Order Modulations | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.2.1.1 Square 16QAM Modulation | New modes for MCS-8 and MCS-9 schemes are introduced that use 16-QAM modulation. With the same payload, 16‑QAM allows lower coding rate (as it enables a higher modem bit rate). Note that 8-PSK, 16-QAM and 32-QAM modulations allow modem bit rates of 1224, 1688 and 2152 per block respectively (excluding the RLC/MAC heade... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.2.1.2 Other 16-ary Modulations | The performance of square 16QAM modulation and Turbo coding schemes (HOMTC) has been evaluated [14], [12], [9]. It has been shown to have much improved performance as compared with EGPRS. However, inherent in using a square 16QAM constellation, there are some practical implementation issues that arise.
- The higher pea... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.2.1.3 32QAM Modulation | 16QAM and circular 16-ary modulation schemes for enhancements in GERAN Evolution [9], [12] and [14]. 32QAM modulation is considered as an additional, and possibly alternative, modulation to 16QAM. 32QAM presents the opportunity to achieve higher peak throughput bit rates, and possibly also more robust channel coding sc... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.2.2 Channel Coding | Much work has been done within the framework of 3GPP RAN standardization with Turbo coding. This chapter takes the turbo coding scheme, and subsequent rate matching as it is used in RAN [15].
The reference configuration used is the existing EGPRS coding schemes MCS-1 to MCS-9. This is compared to the performance of Tur... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.2.3 Symbol Mapping and Interleaving | For non-MCS coding schemes, rectangular interleaving is done across the bursts. For basic block, it is done across 4 bursts. For a doubling of the source code word, interleaving across 8 bursts is used to transmit the block. Bit-wise interleaving is performed. The structure for interleaving over 8 bursts, and the struc... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.2.4 Header Block | The header block would most likely require modification. Since the header is relatively short, a change from Convolutional to Turbo coding is probably not relevant. However, some improvement of the header coding may be required to be properly aligned with the improvement in the performance of the data block reception. ... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.2.5 USF Signaling | The USF signaling to instruct transmission from MS would probably not be affected for a finally selected scheme for 8‑PSK modulation.
For 16-QAM modulation a stronger block code would have to be devised to maintain robustness, and the network would probably be limited for the case that it was required to signal to lega... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.2.6 Link Adaptation | The current mechanism for EGPRS link adaptation is based on BEP reporting. BEP measurements are independent of specific coding scheme used as it essentially estimates the expected uncoded BER. So it is anticipated that the current BEP scheme could be utilized with appropriate modifications to the link adaptation mappin... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.2.7 Incremental Redundancy Combining | The current mechanism for incremental redundancy combining is based on a "family" of MCSs where the members of the family have multiples of a basic payload unit [7]. A modified form of this could be used for Turbo codes MCSs. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.2.8 Multislot Classes | For configurations that interleave a block over 4 bursts, there should be no need for modifications to the Multislot classes. There may be a need for additional multislot class to support signalling of configurations that interleave across dual carriers. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.2.9 Non-core Components | The components described below are not necessary to the core concept of turbo coding within GERAN. However, during the work we observed, particularly for 16-QAM modulation, that the combined effect together with Turbo coding was greater than the sum of the parts. They have therefore been included as sub-components as a... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.2.9.1 Dual Carrier | The dual carrier concept has been described in detail in [7]. We have investigated the improvement in throughput performance that can be obtained by defining a configuration where the turbo-coded RLC block is interleaved between the different carriers (or hopping sequences). This is a specific example of doubling the s... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.2.9.2 MS Receiver Diversity | The potential interdependence of performance enhancements of MSRD and Higher order modulations (with or without Turbo coding) has been investigated in reference [26], when the features are used simultaneously. A particular aim was to assess whether 16QAM/TC would bring a similar improvement also in the case of MSRD, i.... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.3 Modelling Assumptions and Requirements | The impairments include typical imperfections like I/Q modulator/demodulator imbalance, receiver and transmitter synthesizer (phase) noise, frequency error and non-linear characteristics of the power amplifier.
The impairment models used for the simulations are described in SMG2 EDGE workshop contributions from Toulous... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.3.1 Transmitter Impairments | The transmitter impairments that are used in this report are:
- IQ phase imbalance (phase deviation from 90 degrees between I and Q) and IQ gain imbalance (gain difference between I and Q) which are due to the I/Q modulator that produces the analogue baseband signal from the I and Q signals;
- Phase noise due to the sy... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.3.2 Receiver Impairments | The receiver impairments that are used in this report are:
- the I/Q demodulator has IQ phase and gain imbalances as in the transmitter;
- the receiver synthesizer introduces phase noise like the transmitter synthesizer;
- the frequency error can be seen as a constant frequency offset between the reference oscillator a... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.3.3 Equalizer | Set A: A DFSE is used, where 2-taps are handled by MLSE and the rest of the taps are handled by DFE structure. In 8‑PSK modulation the number of states is 8, while in 16-QAM and 32-QAM, the number of states are 16 and 32, respectively.
Set B: The equalizer utilizes the reduced-state sequence estimation (RSSE) with set ... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4 Performance Characterization | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4.1 Implementation Set A | Source: reference [1] and [6].
Implementation set A evaluates the performance of introduction of higher order modulation only using the EGPRS convolutional channel coding. This is now superseded by other implementations where the performance is evaluated for Higher Order Modulation as well as combination of Higher Orde... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4.1.1 Modelling assumptions and requirements | The results are obtained in a co-channel interference limited environment.
A Typical urban channel with 3 km/h mobile speed (TU-3) at 900 MHz carrier frequency is considered.
Single transmit and receive antenna receivers are used.
A linerarised GMSK pulse shaping filter with BT product 0.3 was used.
Blind detection for... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4.1.2 Comparison of BLER Performance | The results indicate that higher order modulations are more sensitive to the impairments compared to 8-PSK modulation for the same back off. However, a good alternative is to increase the back off for QAM modulations to a level that maintains constant impairment. According to our calculations in table 36 the back off v... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4.1.3 Link Performance with Link Adaptation | Link performance for three cases with modified and new modulation schemes according to table 39 including the impact of transmitter/receiver impairments are investigated and compared to existing 8-PSK.
Table 39: Used modulation and bit rate [kbps] for the investigated cases
Case A
Case B
Case C
Case D
Coding scheme
Mod... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4.1.4 System Simulation Results | Systems simulations have been performed with no impairments included. These are not included in this subclause, as we need to consider the impact of impairments comparable with existing HW.
No system simulation results including impairment consideration are available today. To further estimate the throughput gains in t... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4.2 Implementation B | Source: Reference [7]. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4.2.1 Introduction | The impact of the following factors is evaluated:
- Frequency hopping.
- Incremental redundancy.
- Propagation environment.
- RX and TX impairments.
- RRC pulse shaping: Linearized Gaussian pulse (BT=0.3) or Root-raised cosine pulse (rolloff=0.3). |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4.2.2 Basic Link Layer Performance | The purpose of this subclause is to assess the basic link layer performance of the 16-QAM coding schemes. The simulations are carried out without frequency hopping, impairments, and incremental redundancy, which are separately evaluated in the later subclauses. The applied pulse shaping method is linearized Gaussian (B... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4.2.2.1 BER Performance | The raw BER performance is illustrated in the figure 117. As can be seen, 16-QAM is approximately 2 dB to 3 dB less sensitive than 8-PSK.
Figure 117: Raw BER (TU3nFH@900MHz, no impairments) |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4.2.2.2 BLER Performance | The BLER performance of the three highest MCSs is shown in figures 118, 119 and 120 and summarized in table 41.
Figure 118: BLER performance for 8-PSK and 16-QAM without impairments
Figure 119: BLER performance for 8-PSK with and without frequency hopping
Figure 120: BLER performance for 16-QAM with and without frequen... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4.2.2.3 Throughput | The slot-wise throughput is illustrated in figure 121. As can be seen from the figure, 16-QAM brings some improvement to the range of medium and high CIR values, the throughput gain varying between 0 and 10 %. The applied LA switching points are marked with squares.
Figure 121: Throughput (TU3nFH@900MHz, no impairments... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4.2.3 Impact of Frequency Hopping | The results from the frequency hopping simulations are shown in Appendix A and summarized in tables 42 and 43. Due to the lower coding rate and increased interleaving depth, the FH gain is larger for 16-QAM than for 8‑PSK. It is good to note that also MCS-8 and MCS-9 benefit from the frequency hopping in the case of 16... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4.2.4 Impact of Incremental Redundancy | The results from the IR simulations are shown in figure 123, which shows the performance of MCS-9 with and without IR. As can be seen, 8-PSK benefits more from the incremental redundancy than 16-QAM. This is due to the already lower coding rate of 16QAM, which mitigates the gain from the further increases in the redund... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4.2.5 Impact of Propagation Environment | The impact of the propagation environment is illustrated in figure 124. As can be seen, the Doppler effect degrades the throughput of the highest MCSs in the case of fast-moving terminals. In addition, the high delay spread of the HT channel manifests itself as a reduced throughput.
Figure 124: Impact of the propagatio... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4.2.6 Impact of RX and TX Impairments | Given equal coding rates, it is obvious that 16-QAM is more sensitive to the RX and TX impairments than 8-PSK. However, when comparing 8-PSK and 16-QAM with equal payload sizes, the lower operating point of 16-QAM effectively compensates the sensitivity loss due to tighter constellation. This fact can be easily seen fr... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4.2.7 Impact of RRC Pulse Shaping | The performance of 16-QAM could be, in principle, enhanced by replacing the Gaussian pulse with an RRC pulse. The impact of the improved TX filtering is shown in figure 127, which shows the throughput curves for Gaussian and RRC pulse shapes.
Figure 127: Impact of RRC pulse shaping (TU3nFH@900MHz, no impairments, ideal... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4.2.8 Evaluation of Performance | The throughput gain (versus 8-PSK) for different C/I values is shown in figure 129.
Figure 129: Throughput gain from 16-QAM (TU3, typical impairments, ideal LA)
As can be seen, the throughput gain for 16-QAM modulation varies between 0 to 19 % in the case of ideal frequency hopping, and between 0 to 9 % when no hopping... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4.3 Implementation C | Source: Reference [6], [19], [8], [28], [30], [42] and [43]. Note that there are additional details of simulation results in annex B.
Higher order modulations and turbo codes have been proposed as candidates for the GERAN continued evolution feasibility study. The performance gains of 16QAM, 32QAM, turbo codes and the ... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4.3.1 Channel coding | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4.3.1.1 EGPRS | As reference, the regular EGPRS coding schemes MCS-1 to MCS-9 are used, see table 34. The only deviation from table 34 is that Tail bits are viewed as channel coding redundancy. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4.3.1.2 Convolutional Codes with 16QAM | Three 16QAM-modulated coding schemes with convolutional codes are evaluated. These are called MCS-7-16QAM-2, MCS-8-16QAM-2 and MCS-9-16QAM-2 in table 34 and have the same payload size as MCS-7, MCS-8 and MCS-9, respectively. The mother code of EGPRS (R=1/3, k=7) has been used. Uniform (FLO) puncturing has been used for... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4.3.1.3 Turbo Codes with 8-PSK Modulation | Two new 8PSK-modulated coding schemes with turbo codes are evaluated. These are called MTCS-5 and MTCS-6 and have the same payload size as MCS-5 and MCS-6, respectively, as defined in table 44.
The RLC/MAC header, USF and stealing flags have the same channel coding as MCS-5 and MCS-6. The RLC data block has been encode... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4.3.1.4 Turbo Codes with 16QAM Modulation | Three new 16QAM-modulated coding schemes with turbo codes are evaluated, called MTCS-7-16QAM, MTCS‑8‑16QAM and MTCS-9-16QAM, as defined in table 45. These have the same payload size as MCS-7, MCS-8 and MCS-9, respectively.
The RLC/MAC header, USF and stealing flags have the same channel coding as MCS-7-16QAM, MCS-8-16Q... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4.3.2 Modulation | Simulations are run with 8PSK and 16QAM modulation. The 8PSK modulation is according to the definition in 3GPP TS 45.003.
The 16QAM modulation constellation is shown in figure 106. To reduce the Peak-to-AveRage ratio (PAR), the constellation is rotated by /4 radians per symbol. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4.3.3 Pulse Shaping | For both 8PSK and 16QAM, the regular linearized GMSK pulse shape is used. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4.3.4 Link performance Evaluation | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4.3.4.1 Simulation Assumptions | Scenario
The link level scenario is summarized in table 46.
Table 46: Summary of link simulation parameters.
Parameter
Value
Channel profile
TU
MS speed
3 km/h
Frequency hopping
Ideal
Interference
Co-channel
Radio blocks per simulation point
10 000
Impairments
Realistic transmitter and receiver impairment levels have b... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4.3.4.2 Link Level Results | The link level performance at 10 % BLER and 1 % BLER is summarized in tables 47 and 48. Detailed simulation results can be found in annex B.
Table 47: Summary of link level performance @ 10 % BLER
Modulation/coding scheme
C/I @ 10 % BLER [dB]
Total gain (see note) [dB]
Cc/8PSK
Tc/8PSK
Cc/16QAM
Tc/16QAM
5
11.2
10.4
-
-
... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4.3.5 Link-to-system Interface | A two-stage mapping has been used. With this approach, the C/I is mapped to a block error ratio (BLEP) in two stages. In stage one, the model takes burst level C/I samples as input and maps them onto the (raw) Bit Error Probability (BEP) for a burst. In stage two, the BEP samples of one radio block are grouped together... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4.3.6 System Level Results | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4.3.6.1 Simulation Assumptions | Three different sets of modulation and coding schemes are compared. They are summarized in table 49.
Table 49: Evaluated sets of modulation/coding schemes.
Modulation and coding scheme
Set 1
("EGPRS")
Set 2
("16QAM")
Set 3
("16QAM+turbo")
1
MCS-1
MCS-1
MCS-1
2
MCS-2
MCS-2
MCS-2
3
MCS-3
MCS-3
MCS-3
4
MCS-4
MCS-4
MCS-4
5... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4.3.6.2 Results | The three different modulation and coding sets have been investigated in the above described radio network scenarios. Below, the system level results are summarised. Power backoff according to table 50 is included in all simulations. A comparison of performance with and without backoff is presented in subclause 8.4.3.6... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 8.4.3.6.2.1 Scenario 1: 1-reuse with Random Frequency Hopping | Scenario 1, 1-reuse with random frequency hopping, is the tight reuse scenario of the investigation and could for example be users on traffic channels that are tightly planned due to limited spectrum.
Figure 130 shows the average session bit rate for different user percentiles (10th, 50th and 90th percentile). This all... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.