9.4.1 Eight-Carrier HSDPA

Release 8 Dual-Carrier HSDPA (DC-HSDPA) has been widely taken in to use in many HSDPA networks, and it provides data rates of up to 42 Mbps, as introduced in Chapter 3. In Release 9 the solution was extended to cover dual-carrier operation on two different frequency bands, while further carriers were added with up to four carriers in Release 10 and up to eight downlink carriers in Release 11.

The resulting peak data rate (with the use of a 40 MHz spectrum for eight carriers) reaches 336 Mbps with 2 × 2 MIMO. The same peak rate can also be reached with four carriers and 4 × 4 MIMO, but notably the specification does not support simultaneous operation of 4 × 4 MIMO and more than four carriers. Release 11 also enables aggregating non-adjacent carriers on the same frequency band.

Aggregating multiple carriers brings substantial benefits for the end user because any free resources across all carriers can be used flexibly. This provides a dynamic pool of resources, with the multicarrier capable devices able to use any of the carriers with 2 ms resolution, thus allowing an even load with the bursty traffic too, leading to high trunking gain with the dynamic packet sharing over a large number of carriers.

Unlike the single carrier device, the multicarrier capable device may be re-allocated to the carrier or carriers that are experiencing the best propagation and interference conditions every 2 ms, thus maximizing the resulting system throughput.

The gains can be seen in Figure 9.3, which shows the cumulative distribution of the average user throughput and the mean packet call delay for macro cells with an average cell load of 1 Mbps.

The gains depend significantly on the load in the system. If the load is very high and there are a large number of users, then there will be fewer free resources on the other carriers and there is a high likelihood that there will be several users in a good propagation condition on each carrier, thus resulting in lower gains.

For meeting the requirements for the ITU-R IMT-Advanced system, the eight-carrier HSDPA allows aggregation of 40 MHz of spectrum, meeting the requirement for flexible system bandwidth support up to 40 MHz.

9.4.2 Four-Antenna MIMO for HSDPA

The MIMO solution with two transmit and two receiver (2 × 2) antennas was introduced in Release 7, as covered in Chapter 3. To further push the multiantenna transmission capability, MIMO operation with four transmit and four receiver antennas (4 × 4) was introduced. In 4 × 4 MIMO the NodeB uses four transmit antennas and the UE uses four receiving antennas forming four different reception images for each four transmitted signals between the transmitter and receiver, as shown in Figure 9.4. This enables transmission of up to four parallel streams, which doubles the peak data rate compared to 2 × 2 MIMO and also improves the typical cell capacity and user data rates directly benefiting from additional receive antennas, even if four-stream transmission is not possible due to less favorable radio conditions.

This can be seen in Figure 9.5, showing the average macro cell throughput assuming an ideal channel. It can be seen that adding Rx antennas gives more benefits than adding Tx antennas, while the maximum gain is achieved by using four transmit and four receive antennas. In that case, the system will automatically switch between beamforming with four antennas and using the antennas to send up to four parallel streams based on the link quality.

From the perspective of the ITU IMT-Advanced requirements, the 4 × 4 MIMO, when coupled with 64QAM modulation, reaches 17.2 bits/s/Hz peak spectral efficiency, exceeding the IMT-A minimum requirement of 15 bits/s/Hz.

9.4.3 Uplink Beamforming, MIMO and 64QAM

It was also important to improve the performance in the uplink direction, but adding a very large number of carriers was not desirable as that would have degraded the waveform properties and resulted in very difficult power amplifier requirements. Instead, improvements were made for the single carrier transmission (dual-carrier uplink had also been defined earlier) to reach the bits/s/Hz requirement. The following solutions were introduced:

• Uplink beamforming allows uplink dual-antenna transmission to provide better data rate coverage and lower interference from neighboring cells.

• It will also double the peak uplink rate using dual-stream MIMO transmission and triple it when coupled with 64QAM modulation. Analogous to the downlink MIMO, the mobile device uses two transmit paths and antennas to form a complex radio wave pattern in the multiple base station receive antennas, as shown in Figures 9.6 and 9.7. In favorable radio conditions, this yields over 2 dB in the link budget, which translates into up to 30% higher average uplink data rates throughout the cell and up to 40% higher data rates at the cell edge. In another analogy to downlink MIMO, in very good channel conditions and when a high received signal-to-noise ratio is possible, the user equipment may use dual-stream MIMO transmission with two orthogonal beam patterns. This effectively doubles the raw bitrate on the physical layer.

• When further moving from 16QAM to 64QAM modulation, the raw bitrate is tripled together with 2 × 2 MIMO transmission when compared to single stream and 16QAM.

To achieve received signal-to-noise ratios that are high enough to make dual-stream transmission with 64QAM possible over a significantly large area, four or even eight receiver antennas, or a combination of both, may need to be deployed. Yet again this is analogous to what happens in the downlink.

Dual-antenna transmission in the uplink should be viewed as two separate features. First, there is uplink beamforming, which is possible and beneficial in most environments and which helps in the full cell area by providing better uplink data rate coverage. Second is the uplink dual stream MIMO, which is possible only in more limited scenarios and doubles the uplink peak rate.

Introducing 64QAM modulation does not require two transmit antennas, but when aiming for the highest peak rates, it needs to be coupled with uplink MIMO.

Figure 9.8 shows the gains of beamforming on the average and cell edge throughput. It can be seen that gains of up to 20 and 80% respectively can be achieved.

From the perspective of the ITU IMT-Advanced requirements, uplink beamforming helps achieve the average and cell edge performance requirements. The uplink 2 × 2 MIMO, together with 64QAM modulation, achieves 6.9 bits/s/Hz peak spectral efficiency, exceeding the IMT-A minimum requirement of 6.75 bits/s/Hz.

9.4.4 HSPA+ Multiflow

Another feature enabling a better use of resources in cellular systems is HSPA+ Multiflow, which is designed to improve cell edge data rates. Multiflow, as covered in more detail in Chapter 5, enables the transmission of data from multiple cells to a user device at the common cell edge, instead of transmitting the data via a single cell as in HSDPA today. This is illustrated in Figures 9.9 and 9.10 for dual-cell Multiflow operation. With inter-BTS Multiflow the RNC splits the incoming data stream to two independent flows that are transmitted to the UE by two independent NodeBs on the same carrier frequency. With intra-site Multiflow the NodeB MAC layer splits the single data flow coming from the RNC just like in multicarrier HSDPA operation, but the transmission takes place over neighboring sectors of the NodeB on the same carrier frequency, as opposed to multicarrier operation where the data is transmitted on different carrier frequencies, but on the same sector.

Each of the data flows in Multiflow can be scheduled independently, simplifying the concept and enabling simple inter-site deployment without the need for tight network synchronization. The presence of two HSDPA flows from two cells leads to a doubling of the power available for the desired signal at the device, which is used to increase the overall user throughput. For 3GPP Release 11, Multiflow is considered for up to four different flows over two different frequencies, enabling the radio network to send data from two different base stations and up to four different cells to a user device.

Figure 9.11 shows the cumulative distribution of the throughput experienced by the user with and without Multiflow (including both intra-site and inter-site Multiflow devices). At the low values of the cumulative distribution, users at the cell edge gain particular benefit from Multiflow, since they are the most likely to receive transmissions from multiple cells with adequate signal quality.

From the perspective of the ITU IMT-Advanced requirements, HSPA+ Multiflow helps achieve the required performance at the cell edge.

9.4.5 Performance in Different ITU-R Scenarios

The key parameters of the ITU-R deployment environments are described in Table 9.5.

The system performance requirements for a candidate IMT-Advanced radio technology are shown in Table 9.6. The evaluation criteria state that it is sufficient to meet each requirement type in three out of four deployment scenarios to qualify as an IMT-Advanced radio.

The evaluation of Long Term HSPA Evolution performance against these requirements was both a major simulator development effort and a long-lasting simulation campaign. The following figures (Figures 9.12, 9.13, 9.14, and 9.15) show the simulation results against the ITU IMT-Advanced set requirements.

As shown, Long Term HSPA Evolution meets all the IMT-Advanced performance requirements. The results do not include cell edge spectral efficiency features, such as scenario-optimized scheduler or HSPA+ Multiflow, but assume a fairly advanced interference suppressing receiver in the device.

9.4.6 Latency and Handover Interruption Analysis

Table 9.2 listed the ITU-R IMT-Advanced requirements on latency and handover interruption as well as what the UMTS system was already able to achieve. The requirements should be understood as limits that the technology can fulfill, with state-of-the-art implementation and best case deployment.

The user plane latency was analyzed in the same way as was done for LTE-Advanced, looking at the lowest possible air interface delay, and coupling it with HARQ round-trip time averaged with the HARQ retransmission probability. These components are illustrated in Figure 9.16 and Table 9.7.

Control plane latency analysis also followed the practice used in performance evaluation of LTE-Advanced, consisting of the best case latencies from the UE initiating a connection request to the dedicated connection being set up. The latency analysis is detailed in Table 9.8.

Due to the make-before-break nature of soft handover there is no interruption in the intra-frequency handover of WCDMA. HSDPA intra-frequency handover with Multiflow also experiences zero break as first the UE connects to the target cell, and only after that does the UE disconnect from the source cell. Furthermore, with HSPA technology it is possible, with RNC based bicasting and using enhanced HSDPA serving cell change, to achieve zero-break even without Multiflow. In uplink, HSUPA benefits from soft handover and thus by definition there is no break in uplink intra-frequency handovers.

Inter-frequency and inter-band handovers are effectively the same and when executing intra-BTS timing maintained hard handover, even though the handover is break-before-make, the reacquisition of the synchronization is extremely fast. Thus the very basic Release 99 WCDMA standard was already able to meet the handover interruption requirements set much later for IMT-Advanced.