TimeChannel#

class sionna.phy.channel.TimeChannel(channel_model, bandwidth: float, num_time_samples: int, maximum_delay_spread: float = 3e-06, l_min: int | None = None, l_max: int | None = None, normalize_channel: bool = False, return_channel: bool = False, precision: str | None = None, device: str | None = None, **kwargs)[source]#

Bases: sionna.phy.block.Block

Generates channel responses and applies them to channel inputs in the time domain

The channel output consists of num_time_samples + l_max - l_min time samples, as it is the result of filtering the channel input of length num_time_samples with the time-variant channel filter of length l_max - l_min + 1. In the case of a single-input single-output link and given a sequence of channel inputs \(x_0,\cdots,x_{N_B}\), where \(N_B\) is num_time_samples, this layer outputs

\[y_b = \sum_{\ell = L_{\text{min}}}^{L_{\text{max}}} x_{b-\ell} \bar{h}_{b,\ell} + w_b\]

where \(L_{\text{min}}\) corresponds l_min, \(L_{\text{max}}\) to l_max, \(w_b\) to the additive noise, and \(\bar{h}_{b,\ell}\) to the \(\ell^{th}\) tap of the \(b^{th}\) channel sample. This layer outputs \(y_b\) for \(b\) ranging from \(L_{\text{min}}\) to \(N_B + L_{\text{max}} - 1\), and \(x_{b}\) is set to 0 for \(b < 0\) or \(b \geq N_B\). The channel taps \(\bar{h}_{b,\ell}\) are computed assuming a sinc filter is used for pulse shaping and receive filtering. Therefore, given a channel impulse response \((a_{m}(t), \tau_{m}), 0 \leq m \leq M-1\), generated by the channel_model, the channel taps are computed as follows:

\[\bar{h}_{b, \ell} = \sum_{m=0}^{M-1} a_{m}\left(\frac{b}{W}\right) \text{sinc}\left( \ell - W\tau_{m} \right)\]

for \(\ell\) ranging from l_min to l_max, and where \(W\) is the bandwidth.

For multiple-input multiple-output (MIMO) links, the channel output is computed for each antenna of each receiver and by summing over all the antennas of all transmitters.

Parameters:

channel_model – Used channel model
bandwidth (float) – Bandwidth (\(W\)) [Hz]
num_time_samples (int) – Number of time samples forming the channel input (\(N_B\))
maximum_delay_spread (float) – Maximum delay spread [s]. Used to compute the default value of l_max if l_max is set to None. If a value is given for l_max, this parameter is not used. Defaults to 3e-6, which was found to be large enough to include most significant paths with all channel models included in Sionna assuming a nominal delay spread of 100ns.
l_min (int | None) – Smallest time-lag for the discrete complex baseband channel (\(L_{\text{min}}\)). If set to None, defaults to the value given by time_lag_discrete_time_channel().
l_max (int | None) – Largest time-lag for the discrete complex baseband channel (\(L_{\text{max}}\)). If set to None, it is computed from bandwidth and maximum_delay_spread using time_lag_discrete_time_channel(). If it is not set to None, then the parameter maximum_delay_spread is not used.
normalize_channel (bool) – If set to True, the channel is normalized over the block size to ensure unit average energy per time step. Defaults to False.
return_channel (bool) – If set to True, the channel response is returned in addition to the channel output. Defaults to False.
precision (str | None) – Precision used for internal calculations and outputs. If set to None, precision is used.
device (str | None) – Device for computation. If None, device is used.

Inputs:

x – [batch size, num_tx, num_tx_ant, num_time_samples], torch.complex. Channel inputs.
no – None (default) | torch.Tensor, torch.float. Tensor whose shape can be broadcast to the shape of the channel outputs: [batch size, num_rx, num_rx_ant, num_time_samples]. The (optional) noise power no is per complex dimension. If no is a scalar, noise of the same variance will be added to the outputs. If no is a tensor, it must have a shape that can be broadcast to the shape of the channel outputs. This allows, e.g., adding noise of different variance to each example in a batch. If no has a lower rank than the channel outputs, then no will be broadcast to the shape of the channel outputs by adding dummy dimensions after the last axis.

Outputs:

y – [batch size, num_rx, num_rx_ant, num_time_samples + l_max - l_min], torch.complex. Channel outputs. The channel output consists of num_time_samples + l_max - l_min time samples, as it is the result of filtering the channel input of length num_time_samples with the time-variant channel filter of length l_max - l_min + 1.
h_time – [batch size, num_rx, num_rx_ant, num_tx, num_tx_ant, num_time_samples + l_max - l_min, l_max - l_min + 1], torch.complex. (Optional) Channel responses. Returned only if return_channel is set to True. For each batch example, num_time_samples + l_max - l_min time steps of the channel realizations are generated to filter the channel input.

Examples

import torch
from sionna.phy.channel import RayleighBlockFading, TimeChannel

channel_model = RayleighBlockFading(num_rx=1, num_rx_ant=2, num_tx=1, num_tx_ant=4)
channel = TimeChannel(
    channel_model,
    bandwidth=1e6,
    num_time_samples=100,
    l_min=-6,
    l_max=20,
    return_channel=True
)

x = torch.randn(32, 1, 4, 100, dtype=torch.complex64)
y, h_time = channel(x)
print(y.shape)
# torch.Size([32, 1, 2, 126])
print(h_time.shape)
# torch.Size([32, 1, 2, 1, 4, 126, 27])

Attributes

property l_min: int#: Smallest time-lag

property l_max: int#: Largest time-lag

property l_tot: int#: Total number of channel taps

property bandwidth: float#: Bandwidth [Hz]

property num_time_samples: int#: Number of time samples

TimeChannel#

This Page