Qualcomm Technologies, Inc. Not to be used, copied, reproduced, or modified in whole or in part, nor its contents revealed in any manner to others without the express written permission of Qualcomm Technologies, Inc. Qualcomm is a trademark of QUALCOMM Incorporated, registered in the United States and other countries. All QUALCOMM Incorporated trademarks are used with permission. Other product and brand names may be trademarks or registered trademarks of their respective owners. Qualcomm Technologies, Inc. 5775 Morehouse Drive San Diego, CA 92121 U.S.A. 2013 Qualcomm Technologies, Inc. All rights reserved. WCDMA+ Voice and HSPA+ data capacity improvements February 15, 2013 WCDMA+ Voice and HSPA+ data capacity improvements 2013 Qualcomm Technologies, Inc.2 1Introduction A virtuous cycle of user-friendly smartphones & tablets, and improving data rates has led to increasing adoption of data intensive mobile apps and services. This has resulted in an explosion of data traffic which has lead to increasing congestion on HSPA+ mobile networks. Qualcomm has been developing innovative technologies to help network operators address their network capacity needs by means of enhanced network topologies, spectrum aggregation techniques, load balancing techniques and efficiency improvements. This paper describes WCDMA+, a UMTS feature set that significantly improves the efficiency of Release 99 (R99) Circuit Switched (CS) voice calls.In UMTS, voice and data users compete for the same radio resources. As the trend of increasing mobile data traffic is expected to continue well into the future, WCDMA+ can help free up radio network capacity for HSPA+ data services that users increasingly desire.Another consequence of the popularity of these devices is that their large displays, always on applications and high speed modems consume substantial amounts of power resulting in shorter battery life. Aesthetic requirements limit the amount of space available for larger batteries. Any reduction in power consumption for voice calls would help extend battery life, and WCDMA+ addresses this need by reducing modem power consumption and increasing talk time. 2WCDMA+ FeaturesIn recent years, most of the attention in UMTS has focused on improvement to data services through features such as HSDPA, HSUPA, MC-HSPA+, MultiFlow etc., while CS Voice call standards have remained more or less the same since the early days of WCDMA in Release 99. To address this, Qualcomm has proposed WCDMA+ as part of 3GPP Release 12. This section illustrates the feature set that comprises WCDMA+. 2.1Transmit power overhead reduction When designing UMTS R99 CS voice, provision was made for a dedicated pilot in the downlink (DL), which consumes 24% of the DL power budget as shown in Figure 1. The dedicated pilot supports channel estimation at the UE and enables other features like downlink transmit-diversity which did not find major commercial adoption. WCDMA+ improves this design by eliminating the dedicated pilot, and using the Transmit Power Control (TPC) bits for channel estimation. This releases the power budget originally allocated to dedicated pilot for re-use by voice and data channels. The impact on the UE modem design is expected to be small. Figure 1: Today's R99 Downlink Power distribution WCDMA+ Voice and HSPA+ data capacity improvements 2013 Qualcomm Technologies, Inc.3 2.2New frame formatThe current CS voice service in UMTS uses a 20ms frame structure to carry voice frames. This section describes the new frame format that is proposed to make the voice frame transmission more efficient. This new design decreases the transmitter and receiver ON-time, and significantly reduces modem current consumption 2.2.1Downlink frame format On the downlink, WCDMA+ introduces a new frame format that multiplexes two voice calls by splitting the 20ms frame into two 10ms halves. As can be seen in Figure 2, the 20ms Transmission Time Interval (TTI) is split into two 10ms TTIs. Voice traffic from two UEs is multiplexed into the same channel, each occupying alternating 10ms TTIs. The only exception to this is the Transmit Power Control (TPC) bits that are carried even during the TTI that is allocated to the other UE (grayed out blocks).Figure 2: Downlink WCDMA+ frame format This scheme allows for the Node B transmitter to complete sending the frame in 10ms and DTX for the remaining 10ms. The DPCCH also carries the Ack for UL frames received successfully. 2.2.2UL frame format On the uplink (UL), the 20ms TTI is split into two 10ms halves. The UE tries to achieve a 95% transmission success rate in the first 10ms. Should it however fail to do so, the packet is repeated over the next 10ms. Figure 3: UL WCDMA+ frame & control channel format WCDMA+ Voice and HSPA+ data capacity improvements 2013 Qualcomm Technologies, Inc.4 Figure 3 illustrates the new frame structure and the new Ack/Control channel. The new channel carries Transport Format Combination Indicator (TFCI) bits to help early decoding for Frame Early Termination as described in the following section. 2.3Frame Early Termination (FET) As the UE and NodeB experiences dynamic radio conditions, in many cases, they can successfully decode voice frames much earlier than the currently allotted 20ms TTI. This is also facilitated by the error correction ability of the redundancy bits included in the voice frame during the encoding process. However, there is currently no mechanism to take advantage of this phenomenon.WCDMA+ proposes to improve this by enabling the receiver to periodically decode partially received voice frames and sending an acknowledgement upon successful decode. This allows the sender to terminate the transmission early, leading to savings of radio resources, reduced interference and reduced modem power consumption. The new frame format introduced in Sec 2.2 further enhances FET by allowing earlier decoding due to the shorter TTIs.2.3.1Downlink FETFigure 4 illustrates Frame Early Termination in the downlink. As described, the UE stops receiving voice frames after successful decode. A new uplink Ack/control physical channel is introduced to allow the UE to provide feedback to the NodeB of decode success. Figure 4: Downlink FET timeline Figure 5 presents simulation results for early decode success rate for FET in the downlink. These indicate that in as early as 6ms almost 70% of the voice frames are decoded successfully under most channel models. Figure 5:DL FET early decode success rate WCDMA+ Voice and HSPA+ data capacity improvements 2013 Qualcomm Technologies, Inc.5 2.3.2Uplink FET As shown in Figure 6, for FET on the uplink, the ACK for successful decode is carried on the DL control channel and is multiplexed with the DL transmit power control symbols. Once the NodeB successfully decodes a voice frame it acknowledges this to the UE. Figure 6: UL FET timeline When the UE and the Node B have successfully decoded the voice frames, the UE can turn off its receiver and transmitter and the NodeB can stop transmitting to the UE. Figure 7 illustrates results of simulations with the FET mechanism. As observed, UL FET coupled with the new frame format enables successful voice frame decoding with greater than 80% probability as early as 8ms under most channel models. However, should the NodeB fail to decode the frame within the first 10ms, the UE will retransmit it in the following 10ms. Figure 7: FET Early decode success rates (UL) 2.4EVS voice codecAnother major contributor to the network capacity improvement benefit of WCDMA+ is the high quality and more efficient Enhanced Voice Services (EVS) codec that is currently under standardization in 3GPP. Table 1: EVS codec comparison CodecsAvg. Bit RateBandwidthQuality AMR12.2 kbpsNarrowband (0-4 KHz)Todays user experience AMR-WB12.65 kbpsWideband (0-8 KHz)Better than AMR at 12.2 kbps EVS SC-VBR5.9 kbpsWideband (0-8 KHz)Target: equivalent to AMR-WB 12.65k and much better than AMR 12.2 k WCDMA+ Voice and HSPA+ data capacity improvements 2013 Qualcomm Technologies, Inc.6 As shown in Table 1, the Source Controlled-Variable Bit Rate (SC-VBR) variant of the EVS codec family operates at an average of 5.9kbps. Even at about the average data rate of the currently used AMR12.2 codec, the EVS SC-VBR codec is able to provide richer and improved voice quality due to its wideband capability. 3WCDMA+ Performance Benefits As part of its research, Qualcomm has performed link and system simulations to analyze the gains that can be realized from WCDMA+. Figure 8 illustrates the system benefits resulting from the techniques proposed as part of WCDMA+. In todays UMTS systems, a 5MHz carrier can either support 48 R99 CS voice calls using the AMR12.2 codec or carry 5.6Mbps of DL data (Rel 7) in data-only mode, or some combination in between. With WCDMA+, the same 48 CS voice calls will now consume a fraction of the capacity, and the resulting freed up radio resources can now be used to carry about 3.5 Mbps in HSDPA, which is equivalent to about 62% of the data capacity of a 5MHz UMTS data-only carrier. Similar improvement in data capacity (77%) is realizable on the uplink (HSUPA) as well. In addition to the system capacity gains, WCDMA+ results in around 20% - 25% reduction in modem power consumption during CS voice calls leading to a direct improvement in UE talk times. Figure 8: WCDMA+ network data capacity improvement 3.1Components of WCDMA+ gains As described in the previous sections, WCDMA+ achieves its gains through physical layer and voice codec improvements.As can be seen in Figure 9, the physical layer enhancements, including the new frame format and Frame Early Termination, frees up about 37% of the radio resources for HSDPA. The EVS codec improves this further by freeing up an additional 25% of resources for data in the downlink. It is however important to note that benefits from the physical layer enhancements and EVS voice codec are not dependent on each other and can be realized separately. That said, the variable bit rate nature of the EVS codec lends itself more naturally to enhancing the improvements in the presence of FET. Figure 9: WCDMA+ gain components WCDMA+ Voice and HSPA+ data capacity improvements 2013 Qualcomm Technologies, Inc.7 3.2WCDMA+ benefits across operating points The gains modeled in the previous section simulated the system capacity at a load of almost 100% which translates to 48 voice calls per cell. Simulation results in Figure 10 show that during heavy load, system can deliver more than 6x times better data rates with WCDMA+ as compared to an R99 system.Most mobile networks operate at lower loads. For example, even at peak voice hours, call loads may be closer to 25-30 calls per cell. However, as shown in Figure 10, even at these reduced operating points, data capacity improvements are quite substantial. Similar cell throughput and user experience gains are observable on the uplink as well. Figure 10: WCDMA+ benefits across network loads 4Conclusion As users increasingly adopt more advanced mobile devices and data intensive applications, network operators are looking for mechanisms to address this data capacity challenge. Improvements proposed under the WCDMA+ feature set as part of 3GPP Release 12 can provide substantial voice and data capacity improvements to mobile data networks. Benefits include: Improvement to circuit switched (CS) voice efficiency Frees up significant radio network resources (62% on DL and 77% on UL) for mobile data, while maintaining voice capacity Achieves 20%-25% reduction in modem current consumption resulting inUE talk time improvements As mobile networks move to LTE for data services and subsequently to Voice over LTE (VoLTE), both, LTE and HSPA+ networks are likely to co-exist for the foreseeable future. In this multi-mode world, Circuit Switched Fall Back (CSFB) to WCDMA / WCDMA+ will continue to be a primary voice service bearer. Hence, WCDMA+ techniques will continue to be applicable for a significant number of years to come.