Grouping log-likelihood-ratios different multiple-input-multiple-output layers for polar decoding
Various aspects of the present disclosure generally relate to wireless communication. In some aspects, a transmitter may generate multiple encoded bits using a polar encoder. The transmitter may separate the multiple encoded bits into N layers for a multiple-input-multiple-output (MIMO) transmission. The transmitter may select V layer pairs from the N layers, each layer pair of the V layer pairs including a respective first layer and a respective second layer from the N layers. The transmitter may group, for each layer pair of the V layer pairs, a first M symbol bits of the respective first layer and a second M symbol bits of the respective second layer to generate multiple bit pairs. Each bit pair includes a respective first bit of the first M symbol bits and a respective second bit of the second M symbol bits. Numerous other aspects are described.
1 . An apparatus for wireless communication at a transmitter, comprising:
one or more memories; and
one or more processors, coupled to the one or more memories, configured to cause the transmitter to:
generate multiple encoded bits of a data stream using a polar encoder;
separate the multiple encoded bits into N layers for a multiple-input-multiple-output (MIMO) transmission, N being a first integer that is even and greater than zero, the separating being based at least in part on a symbol length of M bits per symbol, M being a second integer;
select V layer pairs from the N layers, each layer pair of the V layer pairs including a respective first layer and a respective second layer from the N layers, V being a third integer;
group, for each layer pair of the V layer pairs, a first M symbol bits of the respective first layer of the layer pair and a second M symbol bits of the respective second layer of the layer pair to generate a MIMO input bit sequence that includes multiple bit pairs, and each bit pair of the multiple bit pairs includes a respective first bit of the first M symbol bits of the respective first layer and a respective second bit of the second M symbol bits of the respective second layer; and
transmit the MIMO transmission using the MIMO input bit sequence.
2 . The apparatus of claim 1 , wherein the one or more processors, to cause the transmitter to group the first M symbol bits of the respective first layer and the second M symbol bits of the respective second layer, are further configured to cause the transmitter to:
group the first M symbol bits of the respective first layer and the second M symbol bits of the respective second layer based at least in part on increasing, at a polar decoder, a log-likelihood-ratio (LLR) magnitude difference between one or more LLR pairs at an input of a polar decoder relative to another LLR pair that, are configured to cause the transmitter to LLRs from a same layer of the N layers, the one or more LLR pairs being associated with the multiple bit pairs.
3 . The apparatus of claim 1 , wherein the one or more processors, to cause the transmitter to select the V layer pairs from the N layers, are further configured to cause the transmitter to:
select the V layer pairs from the N layers based at least in part on an estimated log-likelihood-ratio (LLR) magnitude difference at a polar decoder input.
4 . The apparatus of claim 3 , wherein the one or more processors, to cause the transmitter to select the V layer pairs from the N layers, are further configured to cause the transmitter to:
select the V layer pairs based at least in part on maximizing the estimated LLR magnitude difference.
5 . The apparatus of claim 1 , wherein the multiple encoded bits have a first ordering of bits, and
wherein the MIMO input bit sequence has a second ordering of bits that is different from the first ordering of bits.
6 . The apparatus of claim 1 , wherein the transmitter is a network node, and
wherein the MIMO transmission is a physical downlink shared channel transmission.
7 . The apparatus of claim 1 , wherein the transmitter is a user equipment, and
wherein the MIMO transmission is a physical uplink shared channel transmission.
8 . The apparatus of claim 1 , wherein the one or more processors are further configured to cause the transmitter to:
transmit an indication of a particular mapping sequence,
wherein the one or more processors, to cause the transmitter to group the first M symbol bits of the respective first layer of the layer pair and the second M symbol bits of the respective second layer of the layer pair, are further configured to cause the transmitter to:
group the first M symbol bits of the respective first layer of the layer pair and the second M symbol bits of the respective second layer of the layer pair based at least in part on the particular mapping sequence.
9 . An apparatus for wireless communication at a receiver, comprising:
one or more memories; and
one or more processors, coupled to the one or more memories, configured to cause the receiver to:
receive a multiple-input-multiple-output (MIMO) transmission that includes N MIMO layers, N being a first integer that is even and greater than zero;
demodulate the MIMO transmission to generate a MIMO output log-likelihood-ratio (LLR) sequence that includes N sub-LLR sequences, each sub-LLR sequence of the N sub-LLR sequences being associated with a respective MIMO layer of the MIMO transmission, the MIMO output LLR sequence being based at least in part on a symbol length of M bits per symbol, M being a second integer;
reorder, as at least part of an inverse mapping operation, the MIMO output LLR sequence to generate a decoder input LLR sequence, the reordering being based at least in part on:
the N sub-LLR sequences being partitioned into V sub-LLR sequence pairs, V being a third integer,
each sub-LLR sequence pair of the V sub-LLR sequence pairs including a respective first sub-LLR sequence of the N sub-LLR sequences and a respective second sub-LLR sequence of the N sub-LLR sequences, the respective first sub-LLR sequence being associated with a respective first MIMO layer of the N MIMO layers, the respective second sub-LLR sequence being associated with a respective second MIMO layer of the N MIMO layers, and
the decoder input LLR sequence including multiple LLR pairs, each LLR pair of the multiple LLR pairs being associated with a sub-LLR sequence pair of the V sub-LLR sequence pairs, each LLR pair of the multiple LLR pairs being based at least in part on first M LLRs in the respective first sub-LLR sequence of the sub-LLR sequence pair that are associated with a respective first symbol and second M LLRs in the respective second sub-LLR sequence of the sub-LLR sequence pair that are associated with a respective second symbol, each LLR pair including a first respective LLR of the first M LLRs and a second respective LLR of the second M LLRs; and
generate a decoded bit sequence from a polar decoder by inputting the decoder input LLR sequence to the polar decoder.
10 . The apparatus of claim 9 , wherein, for each sub-LLR sequence pair of the V sub-LLR sequence pairs, the first M LLRs of the respective first sub-LLR sequence and the second M LLRs of the respective second sub-LLR sequence based at least in part on increasing an LLR magnitude difference between LLRs in each LLR pair of the multiple LLR pairs relative to another LLR pair that includes LLRs that are associated with a same MIMO layer.
11 . The apparatus of claim 9 , wherein selection of the V sub-LLR sequence pairs from the N sub-LLR sequences is based at least in part on maximizing an LLR magnitude difference between the first respective LLR of the first M LLRs and the second respective LLR of the second M LLRs.
12 . The apparatus of claim 9 , wherein the MIMO output LLR sequence has a first ordering of LLRs, and
wherein the decoder input LLR sequence has a second ordering of LLRs that is different from the first ordering of bits.
13 . The apparatus of claim 9 , wherein the receiver is a network node, and
wherein the MIMO transmission is a physical uplink shared channel transmission.
14 . The apparatus of claim 9 , wherein the receiver is a user equipment, and
wherein the MIMO transmission is a physical downlink shared channel transmission.
15 . The apparatus of claim 9 , wherein the one or more processors are further configured to cause the receiver to:
receive an indication of a particular mapping sequence; and
derive an inverse mapping sequence using the particular mapping sequence,
wherein the one or more processors, to cause the receiver to reorder the MIMO output LLR sequence, are further configured to cause the receiver to:
reorder the MIMO output LLR sequence to generate the decoder input LLR sequence based at least in part on the inverse mapping sequence.
16 . A method of wireless communication performed by a transmitter, comprising:
generating multiple encoded bits of a data stream using a polar encoder;
separating the multiple encoded bits into N layers for a multiple-input-multiple-output (MIMO) transmission, N being a first integer that is even and greater than zero, the separating being based at least in part on a symbol length of M bits per symbol, M being a second integer;
selecting V layer pairs from the N layers, each layer pair of the V layer pairs including a respective first layer and a respective second layer from the N layers, V being a third integer;
grouping, for each layer pair of the V layer pairs, a first M symbol bits of the respective first layer of the layer pair and a second M symbol bits of the respective second layer of the layer pair to generate a MIMO input bit sequence that includes multiple bit pairs, and each bit pair of the multiple bit pairs includes a respective first bit of the first M symbol bits of the respective first layer and a respective second bit of the second M symbol bits of the respective second layer; and
transmitting the MIMO transmission using the MIMO input bit sequence.
17 . The method of claim 16 , wherein grouping the first M symbol bits of the respective first layer and the second M symbol bits of the respective second layer is based at least in part on increasing, at a polar decoder, a log-likelihood-ratio (LLR) magnitude difference between one or more LLR pairs at an input of a polar decoder relative to another LLR pair that includes LLRs from a same layer of the N layers, the one or more LLR pairs being associated with the multiple bit pairs.
18 . The method of claim 16 , wherein selecting the V layer pairs from the N layers is based at least in part on an estimated log-likelihood-ratio (LLR) magnitude difference at a polar decoder input.
19 . The method of claim 18 , wherein selecting the V layer pairs from the N layers is based at least in part on maximizing the estimated LLR magnitude difference.
20 . The method of claim 16 , wherein the multiple encoded bits have a first ordering of bits, and
wherein the MIMO input bit sequence has a second ordering of bits that is different from the first ordering of bits.
21 . The method of claim 16 , wherein the transmitter is a network node, and
wherein the MIMO transmission is a physical downlink shared channel transmission.
22 . The method of claim 16 , wherein the transmitter is a user equipment, and
wherein the MIMO transmission is a physical uplink shared channel transmission.
23 . The method of claim 16 , further comprising:
transmitting an indication of a particular mapping sequence,
wherein grouping the first M symbol bits of the respective first layer of the layer pair and the second M symbol bits of the respective second layer of the layer pair is based at least in part on the particular mapping sequence.
24 . A method of wireless communication performed by a receiver, comprising:
receiving a multiple-input-multiple-output (MIMO) transmission that includes N MIMO layers, N being a first integer that is even and greater than zero;
demodulating the MIMO transmission to generate a MIMO output log-likelihood-ratio (LLR) sequence that includes N sub-LLR sequences, each sub-LLR sequence of the N sub-LLR sequences being associated with a respective MIMO layer of the MIMO transmission, the MIMO output LLR sequence being based at least in part on a symbol length of M bits per symbol, M being a second integer;
reordering, as at least part of an inverse mapping operation, the MIMO output LLR sequence to generate a decoder input LLR sequence, the reordering being based at least in part on:
the N sub-LLR sequences being partitioned into V sub-LLR sequence pairs, V being a third integer,
each sub-LLR sequence pair of the V sub-LLR sequence pairs including a respective first sub-LLR sequence of the N sub-LLR sequences and a respective second sub-LLR sequence of the N sub-LLR sequences, the respective first sub-LLR sequence being associated with a respective first MIMO layer of the N MIMO layers, the respective second sub-LLR sequence being associated with a respective second MIMO layer of the N MIMO layers, and
the decoder input LLR sequence including multiple LLR pairs, each LLR pair of the multiple LLR pairs being associated with a sub-LLR sequence pair of the V sub-LLR sequence pairs, each LLR pair of the multiple LLR pairs being based at least in part on first M LLRs in the respective first sub-LLR sequence of the sub-LLR sequence pair that are associated with a respective first symbol and second M LLRs in the respective second sub-LLR sequence of the sub-LLR sequence pair that are associated with a respective second symbol, each LLR pair including a first respective LLR of the first M LLRs and a second respective LLR of the second M LLRs; and
generating a decoded bit sequence from a polar decoder by inputting the decoder input LLR sequence to the polar decoder.
25 . The method of claim 24 , wherein, for each sub-LLR sequence pair of the V sub-LLR sequence pairs, the first M LLRs of the respective first sub-LLR sequence and the second M LLRs of the respective second sub-LLR sequence are based at least in part on increasing an LLR magnitude difference between LLRs in each LLR pair of the multiple LLR pairs relative to another LLR pair that includes LLRs that are associated with a same MIMO layer.
26 . The method of claim 25 , wherein selection of the V sub-LLR sequence pairs from the N sub-LLR sequences is based at least in part on maximizing an LLR magnitude difference between the first respective LLR of the first M LLRs and the second respective LLR of the second M LLRs.
27 . The method of claim 24 , wherein the MIMO output LLR sequence has a first ordering of LLRs, and
wherein the decoder input LLR sequence has a second ordering of LLRs that is different from the first ordering of bits.
28 . The method of claim 24 , wherein the receiver is a network node, and
wherein the MIMO transmission is a physical uplink shared channel transmission.
29 . The method of claim 24 , wherein the receiver is a user equipment, and
wherein the MIMO transmission is a physical downlink shared channel transmission.
30 . The method of claim 24 , further comprising:
receiving an indication of a particular mapping sequence; and
deriving an inverse mapping sequence using the particular mapping sequence,
wherein reordering the MIMO output LLR sequence to generate the decoder input LLR sequence is based at least in part on the particular inverse mapping sequence.