PUSCHPrecoder#

class sionna.phy.nr.PUSCHPrecoder(precoding_matrices: List, precision: str | None = None, device: str | None = None, **kwargs)[source]#

Bases: sionna.phy.block.Block

Precodes a batch of modulated symbols mapped onto a resource grid for PUSCH transmissions.

Each transmitter is assumed to have its own precoding matrix.

Parameters:

  • precoding_matrices (List) – List of precoding matrices, one for each transmitter. All precoding matrices must have the same shape. Shape: [num_tx, num_antenna_ports, num_layers].

  • precision (str | None) – Precision used for internal calculations and outputs. If set to None, precision is used.

  • device (str | None) – Device for computation.

Inputs:

inputs – [batch_size, num_tx, num_layers, num_symbols_per_slot, num_subcarriers], torch.complex. Batch of resource grids to be precoded.

Outputs:

x_precoded – [batch_size, num_tx, num_antenna_ports, num_symbols_per_slot, num_subcarriers], torch.complex. Batch of precoded resource grids.

Examples

import torch
import numpy as np
from sionna.phy.nr import PUSCHPrecoder

# Create precoding matrices for 2 transmitters, 2 antenna ports, 1 layer
w1 = np.array([[1], [0]], dtype=complex) / np.sqrt(2)
w2 = np.array([[1], [1]], dtype=complex) / 2
precoding_matrices = [w1, w2]

precoder = PUSCHPrecoder(precoding_matrices)
x = torch.randn(4, 2, 1, 14, 48, dtype=torch.complex64)
y = precoder(x)
print(y.shape)
# torch.Size([4, 2, 2, 14, 48])

Methods

build(input_shape: tuple) → None[source]#

Validate input shape.

Parameters:

input_shape (tuple)