Lte-dci3 and Power adjustment

1.dci3 Format

Although the DCI0, Dci1a, DCI2, dci2a formats can dynamically adjust the transmit power of the uplink pucch and Pusch, these DCI all need to force some RB resources to be assigned to the UE. If some UE does not need to use RB resources, only need to adjust the transmission power of Pusch or PUCCH, then the use of DCI0/1A/2/2A these formats is obviously unreasonable. In order to control the PUCCH or Pusch power of a group of UE via PDCCH, the Protocol introduces DCI3 and 3A formats. These two formats are similar, and this article only describes the DCI3 format.

The purpose of DCI3 is only to be able to dynamically adjust the transmission power of multiple UE's PUCCH or Pusch, not to allocate the upstream or downstream RB resources. When there are more than one UE in the cell at the same time, these UE do not need to allocate RB resources, if you need to uplink power control of these UE, it can be achieved by DCI3 or dci3a.

The DCI3 format is made up of N 2bit TPC command units, "TPC Commands", as shown in table 1.

constituent units	Length (bits)	description
TPC command num ber 1 (user 1)	2	each TPC command unit has a value range of 0~3 to adjust the Pusch or pucch power of a UE, and 1 DCI3 can adjust the uplink power of n UE: when DCI 3 for adjusting the Pusch power, the meaning of the N TPC command unit is the same as the TPC field in DCI0, this time referring to Figure 2, but does not support TPC absolute value adjustment. When DCI3 is used to adjust the pucch power, the meaning of the N TPC command unit is the same as the TPC field in DCI1A, referring to Figure 3. for the use of TPC fields in DCI0 and DCI1A, the relevant post has been described and is not repeated here.
TPC Command number 2 (use 2)	2
...	2
TPC command number N (user N)	2	N=floor (l_format 0/2)    L_format0 represents the effective length of the DCI0 stream under current bandwidth. If Floor (L_FORMAT0/2) < (L_FORMAT0/2), you need to add a bit 0 at the end of the DCI3 stream, because dci3/3a payload size needs to be the same as DCI0 and dci1a.

(table 1) For example, for example, the current 10MHz bandwidth, TDD format and the upper and lower sub-frame configuration is 1, then the payload length of the DCI0 Bitstream is (1+1+11+5+1+2+3+2+1) = 27, so l_format=27,n=13. Because 13< (27/2), you need to add a bit 0 at the end of the DCI3 stream, so the payload size of DCI3 =13*2+1=27, as shown in Figure 1 (which is also used later in this diagram). If the TPC command Number1 corresponds to a 2-bit value equal to (01) = 1, and the current DCI3 is used for Pusch power adjustment (only accumulated adjustment is supported), then by looking at the table in Figure 2 you can know that TPC command NUMBER1 corresponding UE needs to adjust the power value is 0dB, that means that the UE does not need to adjust the Pusch transmit power.

(Fig. 1)

(Figure 2 Table to look for when adjusting Pusch power)

(Figure 3 The table to look for when adjusting the PUCCH power)

2.UE How to determine whether the current pucch adjustment or Pusch adjustment

Since DCI3 can adjust the power of Pusch, can also adjust the power of the PUCCH, then how to differentiate the UE? Here the protocol uses different Rnti values to differentiate. If the current DCI3 is the pucch power adjustment, then DCI3 use Pucch-tpc-rnti to scrambling, and if the Pusch power adjustment, then DCI3 use Pusch-tpc-rnti to scrambling. These two rnti values are similar to C-rnti and Tc-rnti, and are also configured by RRC. TPC-PDCCH-CONFIGPUCCH and Tpc-pdcch-configpusch in cell physicalconfigdedicated are carried to the UE, as shown in Figure 4.

(Fig. 4)

If the current LTE system supports the DCI3 format, then RRC needs to configure Tpc-pdcch-config parameters for the UE in the cell. Since each TPC command unit of DCI3 corresponds to 1 UE, the same pucch-tpc-rnti and Pusch-tpc-rnti values can be configured for several different UE. For example, UE1, UE2, UE3 these three UE, configured to their pusch-tpc-rnti can be the same, the configuration of Pucch-tpc-rnti can also be the same.

3.dci3 is targeted at which UE

The problem of distinguishing between pucch and Pusch power adjustment is solved by different rnti, and the remaining problem is that the TPC command unit in DCI3 is configured for which UE.

In Figure 4 above, you define a field called Tpc-index that can be used to differentiate between UE. If the current ENB is dynamically dispatched in the DCI3 format, the Tpc-index field is 1~15, meaning that a DCI3 can adjust the uplink power for up to 15 UE, while in DCI3A format, the power control can be adjusted for up to 31 UE.

As shown in Figure 1 above, ENB needs to configure different tpc-index for different UE, such as index equal to 1 for UE1, then the 2-bit TPC command unit that is filled in index=1 position in DCI3, is for UE1 assigned If the index=13 is assigned to UE2, then the TPC command unit in DCI3 INDEX=13 is configured for UE2.

4. Example of dynamically adjusting Pusch

We already know that DCI0 can be used to adjust the power of Pusch, dci3/3a can also be used for pusch adjustment, usually in the actual environment we will combine the two to dynamically adjust.

(Figure 5 Pusch power adjustment schematic)

Figure 5 shows a schematic diagram of the Pusch adjustment of the ENB to the ue-a and the ue-b, the actual step is the real value after each dci3/3a adjustment, the dashed line is the target value. As you can see, for ue-a with a faster frequency adjustment, the actual value is closer to the target value, but the faster the adjustment, the greater the load, and for Ue-b, the period of adjustment is longer, the distance from the target is far apart, but the load is also small. When the actual value and the target value distance is too large, you can make an absolute adjustment by DCI0, which is more advantageous to the fast adjustment of power.

Reference:
(1) 3GPP TS 36.212 V9.4.0 (2011-09) multiplexing and channel coding
(2) 3GPP TS 36.213 V9.3.0 (2010-09) Physical layer procedures
(3) 3GPP TS 36.331 V9.18.0 (2014-06) Radio Resource Control (RRC)