Before introducing each transmission mode in detail, first introduce a few concepts.
Before data or signaling is sent to the physical layer, it must be processed by the PDCP, RLC, and MAC layers. But from the perspective of the physical layer, it only receives data from the MAC layer: MACPDU (Protocol Data Unit).
TB
data sent from the MAC layer to the physical layer in the form of transport blocks (transportblock, TB) tissue. One TB corresponds to a data block containing one MACPDU. This data block will be sent within one TTI and is also the unit of HARQ retransmission. If the UE does not support space division multiplexing, one TTI will transmit at most one TB (one TB is equivalent to one codeword). If the UE supports space division multiplexing, one TTI will transmit at most 2 TBs.
Note: (1) One TTI mentioned in the protocol can transmit at most 2 TB, which is from the perspective of the UE. In other words, for a certain UE, in downlink transmission (non-carrier aggregation), the eNodeB can send at most 2 TBs to the UE in one TTI; in uplink transmission, the UE can send at most to the eNodeB in one TTI. 2 TB. From the perspective of eNodeB, it can schedule multiple UEs in one TTI, so eNodeB can transmit or receive more than 2 TBs at the same time in one TTI; (2) Under carrier aggregation, UEs may receive data from multiple cells at the same time , Or send data to multiple cells, so in a TTI may send or receive more than 2 TB, but for a certain cell, the UE can send or receive at most 2 TB in a TTI.
Codeword
A codeword is a data code stream obtained after CRC insertion, code block segmentation, and CRC, channel coding, and rate matching are performed on a TB sent on a TTI. Each codeword corresponds to one TB, so one UE sends at most 2 codewords in one TTI. The codeword can be regarded as a TB with error protection.
Layer
The complex signal (modulation symbol) obtained after scrambling and modulating 1 or 2 codewords is mapped to one or more transmission layers after layer mapping. Each layer corresponds to a valid data stream.
The number of transmission layers, that is, the number of layers, is called "transmission order" or "transmission rank". The transmission rank can be dynamically changed.
The number of layers must be less than or equal to the minimum of the number of transmitting antenna ports and the number of receiving antenna ports.
Section 6.3.3 of 36.211 describes how to perform layer mapping in the following situations.
Single antenna port transmission: There is only one codeword, only layer 1 is used, and the codeword is directly mapped to the layer.
Transmission diversity, with only one codeword, using 2 or 4 antenna ports for transmission, corresponding to layer 2 or layer 4 (but still considered rank=1), all modulation symbols of this codeword are distributed on different layers one by one in a polling manner .
Space division multiplexing: There may be 1 or 2 codewords, which can be mapped to layers 1 to 8.
Note: Although in the LTE specification, transmit diversity is also described as multi-layer transmission, the transmit diversity is still rank=1 (single rank) transmission style.
After layer mapping, the number of modulation symbols contained in each layer is the same and equal to the number of modulation symbols sent on each antenna port. For example: when 2 codewords need to be mapped to layer 3, the number of modulation symbols corresponding to the second codeword (mapped to the second and third layer) is the first codeword (mapped to the first layer) 2 times. This is ensured by selecting the appropriate TB size and rate matching function.
In REL-10, the introduced TM9 supports up to 8 layers of transmission, and its layer mapping method is similar to the previous introduction. The difference is:
(1) 1 TB can be mapped to at most 4 layers;
(2) When the number of layers used is greater than 4, there must be 2 TBs.
The mapping of codewords to layers can be regarded as a process of equally dividing a codeword into N parts, and putting each part into an independent layer. Here N is equal to the number of layers to which a codeword needs to be mapped.
Precoding
Precoding is a process of mapping layers to antenna ports using a precoding matrix. The precoding matrix is an R*P matrix, where R is the transmission rank, that is, the number of transmission layers used, and P is the number of antenna ports.
The codebook is a limited set of available precoding matrices. The UE can only select the appropriate precoding matrix (PMI) from the codebook for reporting. For codebook-based precoding, the eNodeB can only select the appropriate precoding matrix from the codebook and tell the UE the corresponding downlink through DCI The precoding matrix used for PDSCH transmission.
Antenna port
The antenna port is a logical concept. An antenna port can be a physical transmitting antenna or a combination of multiple physical transmitting antennas. In both cases, the UE's receiver is less likely to decompose the signal from the same antenna port, because from the perspective of the terminal, no matter whether the channel is formed by a single physical transmit antenna or multiple physical transmit antennas Combined, the reference signal corresponding to the antenna port defines the antenna port, and the terminal can obtain the channel estimation of the antenna port based on the reference signal. Each antenna port corresponds to a time-frequency resource grid and has its own reference signal. An antenna port is a channel. The terminal needs to perform channel estimation and data demodulation based on the reference signal corresponding to this antenna port.
The relationship between TB, codeword, transmission layer, and antenna port is: TB number = codeword number less than or equal to the number of layers less than or equal to the number of antenna ports.
Reprinted from: https://weibo.com/p/2304186d8537870102w5ja