kaira.channels.AWGNChannel

Inheritance diagram of AWGNChannel — Inheritance diagram for AWGNChannel

class kaira.channels.AWGNChannel(avg_noise_power: float | None = None, snr_db: float | None = None, *args: Any, **kwargs: Any)[source]

Bases: BaseChannel

Additive white Gaussian noise (AWGN) channel for signal transmission.

This channel adds Gaussian noise to the input signal, supporting both real and complex-valued inputs automatically. For complex inputs, noise is added to both real and imaginary components. AWGN channels are fundamental in communication theory and commonly used as a baseline model [Proakis and Salehi, 2007].

Mathematical Model:: y = x + n where n ~ N(0, σ²) for real inputs or n ~ CN(0, σ²) for complex inputs

Parameters:

avg_noise_power (float, optional) – The average noise power σ².
snr_db (float, optional) – SNR in dB (alternative to avg_noise_power).

Example

>>> # For real-valued signals
>>> channel = AWGNChannel(avg_noise_power=0.1)
>>> x_real = torch.ones(10, 1)
>>> y_real = channel(x_real)  # Real noisy output

>>> # For complex-valued signals (same channel works)
>>> x_complex = torch.complex(torch.ones(10, 1), torch.zeros(10, 1))
>>> y_complex = channel(x_complex)  # Complex noisy output

Methods

`__init__`	Initialize the AWGN channel.
`forward`	Apply AWGN to the input signal.
`get_config`	Get a dictionary of the channel's configuration.

Attributes

`avg_noise_power`
`snr_db`

Examples using `kaira.channels.AWGNChannel`

Channel Comparison

Quadrature Amplitude Modulation (QAM)

Original DeepJSCC Model (Bourtsoulatze 2019) with Training

Deep Joint Source-Channel Coding (DeepJSCC) Model - Bourtsoulatze2019 Implementation

Advanced LDPC Code Visualization with Belief Propagation Animation

LDPC Coding and Belief Propagation Decoding

__init__(avg_noise_power: float | None = None, snr_db: float | None = None, *args: Any, **kwargs: Any)[source]

Initialize the AWGN channel.

Parameters:

avg_noise_power (float, optional) – The average noise power σ².
snr_db (float, optional) – SNR in dB (alternative to avg_noise_power).
*args – Variable length argument list passed to the base class.
**kwargs – Arbitrary keyword arguments passed to the base class.

avg_noise_power: float | None

snr_db: float | None

forward(x: Tensor, *args: Any, csi=None, noise=None, **kwargs: Any) → Tensor[source]

Apply AWGN to the input signal.

Parameters:

x (torch.Tensor) – The input tensor.
*args – Additional positional arguments (unused).
csi (Optional[torch.Tensor]) – Channel state information (unused in AWGN).
noise (Optional[torch.Tensor]) – Pre-generated noise tensor. If provided, this noise will be added instead of generating new noise.
**kwargs – Additional keyword arguments (unused).

Returns:

The output tensor with AWGN added.

Return type:

torch.Tensor

get_config() → Dict[str, Any]

Get a dictionary of the channel’s configuration.

This method returns a dictionary containing the channel’s parameters, which can be used to recreate the channel instance.

Returns:: Dictionary of parameter names and values
Return type:: Dict[str, Any]

kaira.channels.AWGNChannel

Examples using kaira.channels.AWGNChannel

Examples using `kaira.channels.AWGNChannel`