Physical Uplink Control Channel, PUCCH carries uplink control information(UCI).
Questions that arises can be listed as:
- What all information is carried in UCI.
- On which subframe is UCI transmitted.
- UE procedure for determining PUCCH assignment.
We’ll find answer for each of these one by one.
What all information is carried in UCI:-
UCI basically carries CQI, SR and HARQ ACK/NACK.
CQI refers to Channel Quality Indicator. CQI is an indicator telling how good the channel quality is. It ranges from 0 to 15, with 15 indicating the best quality. The reason for sending this is that if UE reports high channel quality then network will send a large transport block size. Basically eNb will use a higher modulation scheme and coding rate to increase efficiency. 20 bits per subframe are used to carry CQI.
Granularity of a CQI report can be classified into 3 levels in frequency domain:-
- Wideband CQI– 1 CQI value is reported for the entire downlink channel bandwidth.
- UE selected subband CQI- This is different in periodic and aperiodic reporting(explained below).
For aperiodic reporting, UE divides the bandwidth into a number of subbands, selects best M subbands and reports 1 CQI for the selected subbands.
For periodic CQI, channel bandwidth is divided into bandwidth parts, each of which includes a number of subbands. Ue considers a single bandwidth part each time a ue selected subband report is generated. Ue selects preferred subband from bandwidth part and generates CQI based upon subband.
- Higher layer configured subband CQI– UE reports subband CQI values for each band in a single report.
PUCCH transfers various combinations of SR, HARQ and CQI information based on it’s format.
Formats 1, 1a and 1b transfer HARQ acknowledgments and SR. Formats 2, 2a and 2b transfer HARQ and CSI. Format 3 carries HARQ for carrier aggregation and SR. Following is the summary:-
There are 2 types of reporting modes in time domain:-
- Periodic CQI– UE reports CQI, PMI and RI. CQI is transmitted periodically with reporting period determined by higher layers. Minimum periodicity can be 2 msec. Carri ed by PUCCH or PUSCH. It supports wideband and ue-selected subband CQI. Rate at which UE sends report depends upon CQI-PMI configuration index given to UE in RRC connection setup/reconfiguration message.
UE sends different CSI components periodically, each combination is described as a mode according to following table:-
For each transmission mode, a CSI reporting mode is supported-
- Aperiodic CQI– UE provide detailed reports. It is based on request by eNb therefore, reporting period is given in DCI 0 /DCI 4 or RACH response. Carried by PUSCH. It is not applicable to 1.4MHz band. It supports wideband, ue-selected subband and higher layer configured subband CQI. Similar to periodic reporting, different reporting modes are defined for aperiodic reporting as follows:-
- Different reporting modes are chosen according to different transmission modes.
RRC Connection setup/reconfiguration/re-establishment is used to configure the reporting mode.
RI (Rank Indicator) depicts the number of layers. Basically, it will indicate how well multiple antennas would work together i.e. signals from each antenna would interfere or not.
PMI(Precoding Matrix Indicator) denotes the direction to be given to signals i.e. how to distribute incomming data to antenna ports.
SR and HARQ are explained in different posts.
On which subframe is UCI transmitted:-
When only wideband reporting is configured reports are generated at subframes according to the following formula-
(10 * nf + k – NOFFSET,CQI) mod Npd = 0 (1)
nf = SFN (0 to 1023)
k = subframe no. (0 to 9)
NOFFSET,CQI = reporting offset determined from CQI_PMI configuration index according to table Table 7.2.2-1A specs 36.213.
Npd = reporting period from table.
From cqi-pmi-ConfigIndex, we determine Npd (periodicity)=10 and Noffset,CQI = 1 (with the help of table). Now, we can find the sfn and sf combination that will satisfy above equation.
Now if RI is also configured alongwith wideband CQI, the reporting instances for RI are subframes satisfying:-
(10 * nf + k – NOFFSET,CQI- NOFFSET,RI) mod (Npd *MRI)= 0 (2)
In case of collision of RI and wideband CQI/PMI the wideband CQI/PMI is dropped.
When both wideband and sub-band reporting are configured–
Wideband CQI/PMI reports are sent on subframes satisfying:-
The reporting instances for wideband CQI/PMI and subband CQI are subframes satisfying:-
(10 * nf + k – NOFFSET,CQI) mod (Npd) = 0 (3)
Now, wideband CQI/PMI report has period H*Npd and is reported on subframes satisfying:-
(10 * nf + k – NOFFSET,CQI) mod (H*Npd) = 0 (4)
where,
H = (J*K) +1
J = Number of sub-band parts associated with channel bandwidth (These are predefined in specifications)
K = signalled by eNB from 1 to 4.
NOFFSET,CQI is determined according to table Table 7.2.2-1A specs 36.213.
Sub-band CQI reports are sent during subframes that satisfy eq (1) if the subframe is not already occupied with a wideband report. i.e. between every two consecutive wideband CQI/PMI reports, the remaining J*K ⋅ reporting instances are used in sequence for subband CQI report
RI are reported at different intervals. A table is defined in specifications (Table 7.2.2-1B) for RI configuration index and reporting period.
When RI is configured, reporting interval of RI is-
(10 * nf + k – NOFFSET,CQI- NOFFSET,RI) mod (H*Npd *MRI)= 0 (5)
UE procedure for determining PUCCH assignment:-
The number of RB’s that can be used for PUCCH transmission within a cell is determined by N PUCCHRB (pusch-HoppingOffset). It is broadcasted in SIB.
A single PUCCH transmission occupies 2 RB’s in different time domain. These 2 RB’s are located at the edge of channel bandwidth. i.e. each PUCCH transmission in a subframe comprises of a RB at one end of system bandwidth and an RB at the opposite end of bandwidth. The 2 RB’s are known as PUCCH region. Allocating PUCCH resources at the edges has several advantages:-
- Central RB’s can be used for PUSCH transmission for a single UE. This increases ul data rate. If PUCCH were allocated at the center, PUSCH would be assigned at either end of the control signal to maintain single carrier nature of ul.
- Frequency diversity achieved through frequency hopping is maximized by allowing hopping from 1 edge to another so that if channel conditions are bad at one edge, packets at the other edge can be decoded.
- Out Of Band emissions are reduced, if UE is transmitting on a slot compared to multiple slots.
PUCCH is never transmitted simultaneously (i.e. in the same TTI) with PUSCH from the same UE.
Some terms needs to be understood for PUCCH-
- n(1) RB,
- n(2) RB,
- Cyclic Shift (n(1) CS),
- Orthogonal Sequence(nOC),
- Delta PUCCH Shift (Δ PUCCH SHIFT).
If only HARQ is transmitted-
n(1) PUCCH = nCCE + N(1) PUCCH ; ranges from 0 to 2047, in FDD
For transmitting SR,
n(1) PUCCH,SR is sent in RRC connection reconfiguration information from eNB. a single ue is given multiple PUCCH resources- a standard PUCCH resource and an SR PUCCH resource. Standard PUCCH resource is used when HARQ is transmitted without SR and SR PUCCH resource is used when sending SR with/without HARQ.
N(1) PUCCH is common for all Ue’s , broadcasted in SIB1. It represents number of statically reserved PUCCH resources i.e. some of the first few PUCCH regions that are reserved for format 2/2a/2b or SPS.
nCCE is the number of first CCE used for transmission of corresponding DCI.
For PUCCH format 2, 2a, 2b; n(2) PUCCH determines the PUCCH resource. This is sent by eNB in the RRC connection reconfiguration message.
Cyclic Shift (n(1) CS)-
Multiple UE’s can be multiplexed on 1 PUCCH RB i.e. UE’s can share the PUCCH RB’s but the combination of orthogonal codes and cyclic shift is unique for each UE. This leads to giving different resources to different UE’s on 1 RB.
The number of PUCCH format 1/1a/1b resources per resource block, can be calculated as follows:
NRBPUCCH Format 1 = (NPUCCHRS *12)/ Δ PUCCH SHIFT
where
NPUCCHRS is the number of reference signals on PUCCH format 1/1a/1b (= 3 with normal CP and 2 with extended CP).
Δ PUCCH SHIFT is the cyclic shift difference between two adjacent SR pucch resources using the same orthogonal cover sequence
For normal cyclic prefix,
If Δ PUCCH SHIFT =1, 1 RB can contain 36 resources i.e. 36 UE’s can be multiplexed on 1 PUCCH RB.
If Δ PUCCH SHIFT =2, 1 RB can contain 18 resources i.e. 18 UE’s can be multiplexed on 1 PUCCH RB
If Δ PUCCH SHIFT =3, 1 RB can contain 12 resources i.e. 12 UE’s can be multiplexed on 1 PUCCH RB.
n(1) RB or n(1) PUCCH represents resource used by a specific UE for transmission of PUCCH format 1, 1a 1b.
Following is a UE log snippet for RRC Connection Reconfiguration message-
message c1 : rrcConnectionReconfiguration : {
rrc-TransactionIdentifier 2,
criticalExtensions c1 : rrcConnectionReconfiguration-r8 : {
radioResourceConfigDedicated {
physicalConfigDedicated {
cqi-ReportConfig {
cqi-ReportModeAperiodic rm30,
nomPDSCH-RS-EPRE-Offset 0,
cqi-ReportPeriodic setup : {
cqi-PUCCH-ResourceIndex 12,
cqi-pmi-ConfigIndex 8,
cqi-FormatIndicatorPeriodic widebandCQI : NULL,
ri-ConfigIndex 483,
simultaneousAckNackAndCQI TRUE
}
},
soundingRS-UL-ConfigDedicated setup : {
srs-Bandwidth bw0,
srs-HoppingBandwidth hbw0,
freqDomainPosition 0,
duration TRUE,
srs-ConfigIndex 356,
transmissionComb 0,
cyclicShift cs1
},
schedulingRequestConfig setup : {
sr-PUCCH-ResourceIndex 0,
sr-ConfigIndex 97,
dsr-TransMax n4
}
}
}
very well explained..!!
LikeLiked by 1 person
Very good. Thank you. But how does RRC decide about the reporting period. I mean based on which information it make a decision for a UE?
LikeLike
Hi,
This decision depends on a number of factors.
Although it would be a design and implementation decision, but, to talk about minimal factors, periodicity could vary according to the Bandwidth available, the PUCCH cap, total number of users to be supported in the system, number of users supported in 1 RB, delta-PUCCH shift etc.
Further, SR period may depend on the type of traffic (TCP or UDP) as well as data model (Qos or RR).
I hope it helps.
Thanks,
Neha
LikeLike
Hi Neha,
Well explained !! I have few queries ,hope you can answer them :
1. n(1) PUCCH = nCCE + N(1) PUCCH ; ranges from 0 to 2047, in FDD
Why PUCCH is 0 to 2047 ?
Suppose n(1) PUCCH turns out be any value ,e.g : 12 , then what does ’12’ indicate ??
12th RB ? 12th Sub carrier ? 12th from beginning or end ?
Regards,
Monu
LikeLike
Hi,
I am not sure regarding the limit 2047 for NIPUCCH, will try to figure out the reason.
If n(1) PUCCH turns out be any value ,e.g : 12, then it indicates 12th Resource element.
Thanks,
Neha
LikeLike
Hi, i have one question, When multiple UE sends SR in PUCCH to eNB, how eNB decides which UE has sent the SR, i don’t think C-RNTI is scrambled in the request.
LikeLike
Hi Susanta,
Number of resources reserved for SR is pre-calculated by eNB, from total number of PUCCH resources, number of UEs supported in the system, CQI periodicities, SR periodicities, delta pucch shift etc. This being calculated, eNB sends SR configuration to each UE in RRC connection setup/reconfiguration request. eNB is aware of which UE was supposed to send SR in a particular tti, based on it’s SR configuration.
Thanks,
Neha
LikeLike