Performance Evaluation of Uplink Closed Loop Power Control for LTE System

Performance Evaluation of Uplink Closed Loop Power Control for LTE System Bilal Muhammad and Abbas Mohammed Department of Signal Processing, School of Engineering Blekinge Institute of Technology, Ronneby, Sweden Abstract Uplink power control is a key radio resource management feature in the 3GPP Long Term Evolution (LTE). In order to adapt to changes in the inter-cell interference situation or to correct power amplifier errors, closed-loop adjustments should be applied. In this paper the performance of closed loop power control combined with fractional path loss compensation factor is studied, and an optimal value for the path loss compensation factor is investigated. The closed loop power control with fractional path loss compensation factor is found to improve the system performance in terms of mean bit rate by 68% and utilizes the battery power more effectively. I. INTRODUCTION Power control refers to set output power levels of transmitters, base stations in the downlink and user equipments (UEs) in the uplink. In a multi user environment a number of user shares the same radio resources. A consequence of the limited availability of radio channels in the network is that the same channel has to be assigned to many users. Thus, a signal intended for a certain user will reach other users and introduce interference to their connection, and degrade the quality. A user with very good quality may consider using a lower power and still achieve an acceptable quality, and an additional advantage is that it will disturb other users less, and so their quality will be improved. Power control is essentially to do the same task but in a controlled manner. This paper describes the LTE closed loop power control combined with fractional path loss compensation factor for the Physical Uplink Shared Channel (PUSCH) []. The performance of the fractional path loss compensation [2] combined with open loop power control for 3G Long Term Evolution (LTE) is discussed in [3, 4]. The closed loop power control combined with fractional path loss compensation factor sets the SINR target based on the path loss of the users while the conventional closed loop uses a single SINR target for all the users in a cell. Simulation results will show that the closed loop power control with fractional path loss compensation factor is advantageous compared to closed loop power control with full path loss compensation. The paper is outlined as follows. In Section II an overview of power control is discussed in general and different PUSCH LTE power control schemes are introduced. The proposed closed loop power control algorithm is presented in Section III. In Section IV the simulation assumptions are outlined. The simulations results are presented in Section V followed by conclusions in Section VI. II. POWER CONTROL A. Uplink power control in 3GPP LTE LTE uses single carrier frequency division multiple access (SC-FDMA) as its radio access technology in the uplink. Usage of an orthogonal transmission scheme eliminates mutual interference between users in the same cell (intra-cell interference) and near-far problem as of typical CDMA systems. However, since transmission in the neighboring cell is not orthogonal, there is interference between users in the neighboring cells (inter-cell interference) that ultimately limits the system performance in terms of capacity. In order to maximize the spectral efficiency, 3GPP LTE is designed for frequency reuse [5] both for downlink and uplink, which means that all cells in the network use the same frequency bands. Thus, with frequency reuse, both data and control channels are sensitive to inter-cell interference. The cell edge performance and the capacity of a cell site can be limited by the inter-cell interference. Therefore, the role of closed loop power control becomes decisive to provide the required SINR (signal-to-interference-and-noise ratio) to maintain an acceptable level of communication between the enb and the UE while at the same time controlling interference caused to neighboring cells. In addition, battery power is a scarce resource for portable devices such as notebooks, ultra-portables, gaming devices and video cameras. In the coming years these devices will operate over mobile broadband technology such as LTE. Therefore, in order to minimize consumption of battery power and use the available power efficiently, power control can be helpful. The 3GPP specifications [4] defines the setting of the UE transmit power for PUSCH by 978--4244-255-/9/$25. 29 IEEE

PPUSCH = min{ P max, log M + P + α PL+ δmcs + f( Δ i)} () - P max is the maximum allowed transmit power. It depends on the UE power class. - M is the number of physical resource blocks (PRB). - P is cell/ue specific parameter. It is used to control SNR target and is signaled by the radio resource control (RRC). In this paper, it is assumed that P is cell-specific. - α is the path loss compensation factor. It is a 3-bit cell specific parameter in the range [-] signaled by RRC. - PL is the downlink path loss estimate. It is calculated in the UE based on the reference symbol received power (RSRP). - δ mcs is cell/ue specific modulation and coding scheme defined in the 3GPP specifications for LTE. - f (Δ i ) is UE specific. i Δ is a closed loop correction value and f is a function that permits to use accumulate or absolute correction value. The parameter P is calculated as [6] P = α ( SNR + Pn ) + ( α)( P max log M ) [dbm] (2) SNR is the open loop target SNR (signal-to-noise ratio) P n is the noise power per PRB. M defines the number of PRBs for which the SNR target is reached with full power. It is set to for simplicity. The power requirement to transmit a single resource block can be calculated as PSDTx = P + α PL dbm / PRB (3) where PSD is the power spectral density. B. Conventional power control scheme This scheme is widely used in cellular systems which are not using orthogonal transmission scheme in the uplink, such as conventional CDMA based systems. One of the advantages of this power control scheme is that it combats the near-far problem typically experienced by CDMA systems, since it equalizes the power of all UEs received at the base station. Allowing full compensation of the path loss (i.e., α = ), P is given by = + (4) P SNR Pn dbm Taking into account the path loss, the received PSD at the enb is then given by PSDRx = ( SNR+ Pn) = P [ dbm/prb] (5) It is clear from eq. (5) that the received PSD at the enb is equal to P, thus illustrating that the conventional power control scheme steers all users with equal power spectral density. C. Fractional power control scheme The fractional power control (FPC) scheme allows the users to be received with variable PSD depending on their path loss; that is the user with good radio conditions will be received with high PSD and vice versa. Setting the path loss compensation factor in the range < α <, P can be stated as P = α ( SNR + Pn ) + ( α)( P max) [dbm] (6) The received PSD in this case is where PL is the path loss. PSDRx = P + PL( α ) dbm/prb (7) Attention is drawn here by comparing eq. (5) and eq. (7), since the received PSD in the case of conventional power control scheme results in P while in the fractional power control scheme it also have the additional term PL(α-). Since both P and α are cell specific parameters broadcasted towards the UEs by the enb (same for all the UEs), then PL is the key factor in eq. (7) that allows users to be received with different power spectral density. The choice of the value of α leads to conventional or fractional power control scheme. Both the conventional and fractional open loop power control schemes are discussed in [4]. D. Closed loop power control scheme Closed loop power control is the ability of UE to adjust the uplink transmit power in accordance with the closed loop correction values which is also known as transmit power control (TPC) commands. In LTE closed loop power control system, the uplink receiver at the enb estimates the SINR of the received signal and compares it with the desired SINR target value. If the received SINR is below the SINR target, a TPC command is transmitted to the UE to request for an increase in the transmitter power. Otherwise, the TPC command will request for a decrease in transmitter power. The LTE closed loop power control mechanism operates around the open loop point of operation. The UE adjusts its uplink transmission power based on the TPC commands it receives from the enb when the uplink power setting is performed at the UE using open loop power control. In this paper, accumulate correction value [] is considered to correct the UE uplink power, which is calculated from f() i = f( i ) + δ ( i K ) db (8) PUSCH PUSCH

where δ PUSCH is UE specific correction value, which is also referred as TPC command. TPC commands [-,,, 3] db are used during the simulations. f () = and K PUSCH = 5 TTI s (transmission time intervals). The conventional closed loop power control scheme keeps the same SINR target setting for all the users in a cell and uses α =. III. PROPOSED CLOSED LOOP POWER CONTROL ALGORITHM In conventional closed loop power control the SINR target setting is the same for all users and is a trade-off between the cell-edge and mean bit rate; that is a high SINR target results in high mean user bit rate but lower cell-edge bit rate, while lower SINR target leads to low mean and high cell-edge bit rate. In contrast to the conventional closed loop power control, the proposed scheme sets SINR target for each user based on their path loss which allows users with good radio conditions to achieve better SINR and at the same time providing better cell-edge bit rate. A. Setting the SINR target based on the path loss of users A mathematical expression needs to be derived to set the closed loop SINR target based on the path loss of users while keeping the baseline SINR target for those users which are using full power (P max ). Therefore, a relation is formed between the received SINR and path loss (PL) of the users with the aid of an illustration shown in Fig.. In the figure, PL max is the maximum path loss, at which users start using P PUSCH = P max. PL < PL max is the path loss of any arbitrary user, and SINRtarget is the closed loop baseline SINR target to start with. SINRtarget is the SINR target based on the path loss and α is the path loss compensation factor while m is the slope and is given by α, and IN is the interference and noise power in dbm. The knee point in Fig. is denoted by PL max where users use P max at this point and beyond. The users at PL < PL max are experiencing relatively better radio conditions than users at PL > PL max, and so it is desirable that users should take advantage of their good radio conditions. Fig. : Illustration of setting the SINR target based on the path loss of users. Using Fig. and the existing information discussed above, then the required mathematical equation that provides SINR target based on the path loss of users can be obtained using the slope of the line and is given by m = Δ (9) Δ Y X Δ Y = SINRtarget - SINRtarget [ db ] () By using eq. (), PL can be defined as Δ X = PL PL max [ db ] () PL = [ db] (2) α P M P f Δ { PUSCH log ( i)} The PL involves P PUSCH as can be clearly seen in eq. (2). In the real world, however, the enb uses power headroom report (P h ) [7] received by enb from the UE in order to find the path loss of each user. Thus, PL can be rewritten as PL = α P M P f Δ { h log ( i)} where P h = P PUSCH = P max when PL = PL max. [ db] (3) It is worthwhile to note that power headroom reporting is not considered in this paper, thus P PUSCH is used instead of P h to calculate the path loss. Power headroom reporting can be considered as an extension of this paper. By using eqs. (9-2), the SINR target based on the path loss is given by ( α ) ( PL PL max) + SINRtarge t, PL < PL max SINRtarget = [db] SINRtarget, PL PL max (4) In eq. (4), users at PL > PL max will use SINRtarget = SINRtarget, indicating that there is no increase in the SINR target for users which are already using P PUSCH = P max. Furthermore, using α = turns the designed closed loop scheme into conventional closed loop power control implying that the SINR target setting is independent of the path loss.. IV. SIMULATION ASSUMPTIONS AND MODEL Dynamic simulations have been used. The terminals having the velocity of 3 km/h are randomly positioned in the system area, and the radio channel between each base station and terminal antenna pair is calculated according to the propagation and fading models. The simulator used the raybased 3GPP Spatial Channel Model Extension (SCME) [8] to model the multipath fading propagation in the system.

No measurement or power setting error is included. The simulation parameters are listed in Table. Traffic Models User distribution Uniform Terminal speed 3 km/h Data generation Simple upload traffic model Radio Network Models Distance L = 35.3+37.6*log(d), d = distance in meters attenuation Shadow fading Log-normal, 8dB standard deviation Multipath fading Cell layout Cell radius System Models Spectrum allocation Max UE output power Max antenna gain Modulation and Coding schemes Scheduling Receiver Table : Default simulation parameter SCM, Suburban macro Hexagonal grid, 3-sector sites, 2 sectors in total 67m (5m inter-site distance) MHz (5 resource blocks) 8kHz ( resource block) 25mW into antenna 5dBi QPSK and 6QAM, turbo coding Round robin MMSE [9] with 2-branch receive diversity The LTE closed loop power control operates around open loop point of operation. Thus, the simulator sets the open loop power control based on the SINR target and the power is corrected using TPC commands which are issue based on the difference between SINR target and estimated received SINR. Based on transmit power, channel realizations, modulation scheme, and the active interferers, an SINR is calculated for each link including both intra- and inter-cell interference. The simple upload traffic model is used in this paper rather than the full buffer traffic model. In full buffer model, neither a user leaves due to hang up, nor does a new user arrives, since each user buffer is filled with infinite data, and the user will not leave until it transmits all the data. The simple upload traffic model is designed in a way such that the users can have limited data in their buffers, and so a user leaves when it transmits the data and new users are added in the system. It provides the ease to define the user upload file size and mean bearer bit rate. The mean bearer bit rate along with the offered cell throughput defines the total number of users in the system. Moreover, the simple upload buffer model also allows inclusion of the effect of queuing delay when calculating the user bit rate. The queuing delay reflects more realistic results and provides a better scale for performance comparison in choosing the optimal value of α. For different values of α, the 5 th percentile and mean user throughput is calculated by taking the effect of queuing delay into account. V. SIMULATION RESULTS In this section we present the simulation results using the simple upload traffic model. Simple upload buffer model provides more realistic results and a better scale for performance analysis. Each value of α is investigated for each closed loop SINR target from a set of SINR targets. The criterion that selects optimal value of α for a given SINR target is optimized for cell-edge bit rate; that is the value of α which gives the best cell edge performance for a given SINR target will be chosen. A. Investigating optimal value of the path loss compensation factor Fig. 2 shows the cell-edge and mean bit rate for different values of α. It is clear that α = results in low cell-edge bit rate when compared to α =.8. This is due to the relatively high interference since more users staying in the system for longer time. The users stay for longer time because they cannot transmit their data quickly due to the low mean bit rate. It is also evident from Fig. 2 that the closed loop with full compensation results in lower mean bit rate. The longer the users stay the more will be the users in the system, which leads to high queuing delay and ultimately the interference will rise. If we look at α =.7, it can be seen that it results in higher mean bit rate but degrades in the cell-edge performance due to the use of more power which results in the rise in interference level, and thus the users at the cell-edge will be affected. The path loss compensation factor close to results in increase in the uplink power. Therefore, α =.8 is a better value for fractional path loss compensation factor, as it results in the best cell-edge and better mean bit rate. User bit rate [Mbps].9.8.7.6.5.4.3.2. Investigation of α using simple upload traffic model cell-edge bit rate mean bit rate.7.75.8.85.9.95 α Fig. 2: Cell-edge and mean bit rate performance for different values of α for the closed loop power control using simple upload traffic model. The parameter settings are optimized for the best cell-edge bit rate. B. Performance analysis in terms of uplink received SINR of the closed loop power control using α =.8 Fig. 3 shows the CDF plots of the uplink average received SINR using simple upload traffic model. It can be seen from

this figure that the closed loop power control with full compensation steers all users to achieve equal uplink average received SINR, as seen in both the 5 th percentile users and the users close to the base station (i.e., users with good radio conditions). In this case all users get equal received SINR since they all aim to achieve the same baseline SINR target. In contrast to full compensation, the closed loop power control with fractional compensation keeps the baseline SINR target for the worst users and at the same time increases the baseline SINR target based for the users with good radio conditions based on their path loss, where low path loss results in high increase in the SINR target. The effect of SINR target setting based on path loss of the users is clearly evident from Fig. 3, which shows that the better the radio conditions the higher is the average received SINR. figure that the overall power utilization is roughly the same using the closed loop power control with full and fractional compensation. However, the closed loop power control using α =.8 utilizes battery power more efficiently as it provides better system performance in terms of cell-edge and mean bit rate than the closed loop power control with full compensation. 9 8 7 6 5 4 3 2 Power utilization 9 Performance in terms of received SINR, SINR target : db α:.8 α:.5..5.2.25 Power [W] 8 7 6 5 4 3 2 α:., Ideal α:.8, Ideal -5 - -5 5 5 Uplink Average Received SINR [db] Fig. 3: CDF plot of the uplink average received SINR. C. Performance analysis in terms of user bit rate of the closed loop power control using α =.8 Fig. 4 shows the performance gain of the closed loop power control using α =.8 in both the mean and cell-edge bit rate. The mean bit rate is improved by 68% and at the same time providing better cell-edge performance than α =. 9 8 7 6 5 4 3 2 Performance in terms of user bit rate, SINR target: db α:, Ideal α:.8, Ideal..2.3.4.5.6.7.8.9 User bit rate [Mbps] Fig 4. CDF plot of the user bit rate comparing α =.8 and. D. Performance analysis in terms of power utilzation of the closed loop power control using α =.8 Fig. 5 show the UE power utilization with full and fractional compensation using α =.8. It can be clearly seen from this Fig. 5: UEs power utilization using the closed loop power control with full and fractional compensation. VI. CONCLUSIONS In this paper the performance of closed loop power control combined with fractional path loss compensation factor is studied, and an optimal value for the path loss compensation factor is investigated. The simulation results have shown that conventional closed loop power control can be replaced by the closed loop power control combined with fractional path loss compensation factor, thus improving the system performance in terms of the mean and cell-edge bit rates. The performance in terms of mean bit rate could be improved by 68% for a given SINR target. Moreover, the closed loop combined with fractional path loss compensation factor utilizes the battery power more efficiently. REFERENCES [] 3GPP E-UTRA Physical layer procedures, TS 36.23 V8.. [2] J.F. Whitehead, Signal-Level-Based Dynamic Power Control for Cochannel Interference Management, VTC 993. [3] C. Castellanos, D. Villa, C. Rosa, K. Pedersen, F. Calabrese, P. Michaelsen and J. Michel, Performance of Uplink Fractional Power Control in UTRAN LTE, IEEE Vehicular Technology Conference, VTC Spring 28. [4] Simonsson, A.; Furuskar, A.; Uplink Power Control in LTE - Overview and Performance, Subtitle: Principles and Benefits of Utilizing rather than Compensating for SINR Variations, IEEE Vehicular Technology Conference, VTC Fall 28. [5] 3GPP TS 36.23 V8.2. E-UTRA Physical layer procedures [6] R-7485 Uplink Power Control for E-UTRA Range and Representation of P. [7] R4-862 LS on power headroom reporting. [8] D. Baum et.al., An Interim Channel Model for 4G Systems, Extending the 3GPP Spatial Channel Model (SCM), IEEE Vehicular Technology Conference, VTC Spring 25. [9] J. H. Winters, Optimum Combining in Digital Mobile Radio with Cochannel Interference, IEEE Journal on Selected Areas in Communications, Vol. SAC-2, No. 4, July 984.