LTE downlink Physical layer transport mechanism (5) selection and dci1a of-DCI format

The PDCCH channel transmits control information associated with the physically up-and-down shared channel (PUSCH, Pdsch), the DCI information (downlink control information), These DCI messages include information such as RB resource allocation, modulation MCS, Harq-id, and many more related content. The terminal only correctly decodes the DCI information in order to properly process Pdsch data or Pusch data.

Different DCI information, its purpose can be different, for example, there is a DCI for the downlink RB resource allocation, there is a DCI for the uplink RB resource allocation, there is a DCI for uplink power control adjustment, there are special for downstream dual-stream space division multiplexing DCI. The protocols classify these DCI and differentiate them in different DCI formats.

In the R9 version of the agreement, there are a total of the following definitions of Dci:dci0, DCI1, dci1a, dci1b, dci1c, dci1d, DCI2, dci2a, dci2b, DCI3, dci3a. Among them,DCI0, DCI3, dci3a is the DCI type related to uplink Pusch or PUCCH, while DCI1, Dci1a, dci1b, dci1c, dci1d, DCI2, DCI2A, DCI2B, is the DCI type for downlink Pdsch .

1. Which DCI format to use in different scenarios

With so many DCI formats, what specific DCI format is required for a particular PDCCH channel?

if Pdsch is scrambling by Si-rnti , then the DCI format can use DCI1A or dci1c, depending on the implementation of the ENB side to decide which format to use, such as the following table. As can be seen from the table, this time the CCE is only distributed in public search space, the end of the blind inspection is only needed in public space.

if Pdsch is scrambling by P-rnti , then the DCI format can use DCI1A or dci1c, as in the following table.

if Pdsch is scrambling by Ra-rnti , then the DCI format can use DCI1A or dci1c, as in the following table.

if Pdsch is scrambling by Tc-rnti , then the DCI format can use DCI1A or dci1c, as in the following table.

If Pdsch is scrambling by C-rnti , then the DCI format needs to be considered in the TM mode ( transmissionmode parameter) of the current terminal, as in the following table. At this point, the CCE search space is no longer confined to public search space.

if Pdsch is disturbed by Sps-rnti , then the DCI format also needs to synthesize the current terminal's TM mode considerations, such as the following table.

2. Usage scenarios for different rnti

The different rnti are mentioned above, and the scenarios used for each rnti are shown in the table below. For example, when sending a paging message or an SI change notification message, it is necessary to use P-rnti scrambling Pdsch data, and tc-rnti scrambling is required when sending MSG3 or conflict resolution is required.

Different Rnti, its value range is not the same, such as the following table.

High and low position problem of 3.DCI bitstream

The DCI Stream is transmitted in bits (bit) , which means that the same field (such as the RB assignment field, MCS field, Ndi field, and so on) may span multiple bytes, This involves the need to be aware of the high and low level issues when stitching multiple bits into a full field. For example, if a field in the DCI occupies 5 bits, and 1 bits are allocated in a low byte, and the other 4 bits are allocated in a high byte, then how to successfully stitch the 5 bits into a complete field is clear.

To illustrate this, the protocol first contracts thatall the fields in the DCI stream are mapped sequentially to A0,A1,..., an, the 1th field is mapped from A0 , and the bits in the stream are populated sequentially from the low byte to the high byte, as shown.

Second, if a field includes more than one bit, the MSB corresponds to the lowest bit of the bit stream . For example, if the stream of the first 12 bits of a DCI is binary 100111110111, the value of field 1 = 1, the value of field 2 = 0, the value of field 3 is =bin (01111) = 15, and the value of field 4 is =bin (10111) = 23. ENB side at the time of encoding, the value of the first byte of the DCI =bin (10011111) =0x9f, the value of the second byte =bin (01110000) =0x70 (assuming A11 after the bit is 0).

Definitions for each field in the 4.DCI1A format

This article describes the more DCI1A formats used in the downstream dispatch allocation. PDCCH channel bearer of the DCI1A format of the code stream, it has 2 kinds of functions: the first role, is also the most commonly used role, is the bearer of the current sub-frame pdsch scheduling control information ; the second function is the network side The PDCCH Order is initiated by DCI1A and the non-competitive random access process is triggered . For non-competitive random access related content, please refer to the LTE-TDD random access process (6)-sub-scenario description of competitive access and non-competitive access process.

If DCI1A is used as PDCCH Order to trigger a non-competitive random access process, then the bit bits in each of the DCI fields are defined as follows:

- dci0/dci1a format indicates bit –1bit. When the bit value = 0, the current DCI format is DCI0, when the bit value = 1, indicates that the current DCI format is dci1a. - centralized or distributed VRB allocation indicator bit . –1 bit. Fixed 0. - RB Resource Allocation field – [log2 (n_dl_rb* (n_dl_rb+ 1)/2)] to take the whole bits up . Sometimes abbreviated to the RIV field. The value of all bit bits is fixed at 1. For example: The current downlink bandwidth is 1.4MHz, then n_dl_rb= 6, so [log2 (n_dl_rb* (n_dl_rb+ 1)/2)] rounding up =[log2 (6* (6+1 )/2)] [Rounding up =[LOG2 (21)] rounding up =[4.39] rounding up = 5. As a result, 5 bits of the field are fixed to fill 1. - Preamble Indexes Preambleindex –6 bits. This parameter is configured by the ENB side of the RRC, and the terminal launches the non-competitive random access, which needs to be filled in with the preamble index in Prach. - Prach Mask index Prach mask, index –4 bits. This parameter is also configured by the ENB side of the RRC, which determines the location of the Prach channel, with reference to the LTE-TDD random access process (4)-riv parsing and preamble resource selection. - The remaining bits bits are filled in 0. For a specific downlink bandwidth, the DCI1A has a fixed bit length, such as LTE-TDD format and downlink bandwidth =20mhz, dci1a length is 31bits (the specific calculation process is described below), then when dci1a as Pdcch order, At this time the valid field only used (1bit+1bit+13bits+6bits+4bits) =25bits, and (31-25) =6bits not used, at this time need to fill the end of the 6bits all 0.

If DCI1A is not used as a non-competitive access, that is not used as a pdcch Order, but as a dispatch information for Pdsch, then the bit bits of the DCI fields are defined as follows:

- dci0/dci1a format indicates bit –1bit. When the bit value = 0, the current DCI format is DCI0, when the bit value = 1, indicates that the current DCI format is dci1a. - centralized or distributed VRB allocation indicator bit . –1 bit. Fill in 0 if it is a centralized VRB assignment. - RB Resource Allocation field – [log2 (n_dl_rb* (n_dl_rb+ 1)/2)] , rounding bits up, calculating bit bit and PDCCH Order is the same. It is necessary to note that DCI1A only supports the continuous distribution of RB scenarios, so at this time the RB resource allocation involves 2 questions, the first is where the Pdsch data occupies the starting position of the RB, and the second is the length of the RB occupied. To solve both of these problems, you need to encode both information in this one Riv field. The agreement is such as to stipulate the meaning of Riv: For example, a Pdsch corresponding to the RB start position =4,rb length = 8, the downlink bandwidth is 20MHz, then the formula l_crbs=8,n_dl_rb=100,rb_start=4, at this time: L_crbs-1=8-1=7, (N_DL_RB/2) takes the whole =50>7 down. So at this time ENB fill in the riv=100x (8-1) +4=704. - MCS field –5bits. Encoding and modulation of data in Pdsch. The range is 0~28. - harq-id number –3 bits (FDD), 4 bits (TDD). Because FDD has a fixed number of Harq of 8, 3 bits is sufficient. - ndi indicator –1 bit. The terminal determines whether the current Pdsch data is new or retransmission based on whether the NDI has a reversal. If reversed, it represents a new pass, and if there is no reversal, it is a retransmission. - RV Redundant version –2 bits. If it is retransmission, generally according to 0, 2, 3, 1, 0 of the regular transmission, but the net side in the actual implementation, in some scenarios, may not necessarily exactly in accordance with this order. - TPC command for PUCCH –2 bits. There are 2 different scenarios to distinguish: (1) if it is Ra-rnti, P-rnti, Si-rnti scrambling: At this time, 2bit TPC only use this bit of LSB, if lsb=0, it means n_1a_rb=2, if lsb=1, it means n_1a_rb=3 . However, it should be noted that this n_1a_rb=2 or 3 does not mean that the corresponding Pdsch can only use 2 RB or 3 RB. For example ENB issued SIB,LSB (TPC) = 0, at this time n_1a_prb=2, but Pdsch use of RB is based on the previous description of the "RB Resource Allocation field" to decode, that SIB can use 4 RB or more RB. However, regardless of the number of RB used by Pdsch, the TBS found in the 36213-7.1.7.2.1 table must be available in the rb=2 or rb=3 column TBS. For example: Send Sib dci1a, decode riv field to get rb=4, decode mcs field = 1, decode TPC field = 3. Depending on the RB and MCS fields, tbs=144bits can be obtained, and this 144bits should also be found in the Rb=3 column of the table. Actually check the 7.1.7.2.1 table, you can find that 144bits is indeed in rb=3 this column exists, and at this time mcs=2. (2) if it is crnti and other rnti scrambling : Then this 2bits represents the TPC adjustment value of the PUCCH channel. The range of 2bit is 0~3, and different values represent different power adjustment values, as shown in the following table. If the terminal decoding to the tpc=1, then the pucch power adjustment amplitude =0db, that is, the next transmission power of the PUCCH channel is not adjusted; if tpc=2, the next transmission power of the PUCCH channel is adjusted upward 1dB. - dai field –2 bits. This field is only used in LTE-TDD and does not have this field in FDD. For TDD, this Dai field is also valid only when the upper and lower sub-frame configuration is 1~6. We know that the transmission of data in the empty port is very likely to be lost, for Pdsch or Pusch, can be guaranteed by Harq and ARQ, but for the control channel of the DCI, there is no handshake mechanism, so the network side after the issue of a pdcch, It is not clear whether the terminal has received this DCI message, or that the terminal does not know that it has missed the DCI. Then the introduction of this Dai field can be used to determine whether there is a loss of DCI, such as the terminal decoded to the same Harq process of 2 dci1a, wherein the DAI value is 1 and 3, then the terminal can be aware that the HARQ process dai=2 the DCI1A did not detect, was lost. More detailed information about Dai, in the follow-up on the downstream HARQ response to the specific discussion.

If the current downlink bandwidth is 20MHz, using the LTE-TDD format, then the dci1a bit length = (1+1+13+5+4+1+2+2+2) =31bits.

Reference documents:
(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 36.321 V9.6.0 (2012-03) Medium Access Control (MAC) protocol specification

LTE downlink Physical layer transport mechanism (5)-DCI format selection and dci1a