Channels

Define channel models for communication systems. Custom channels can be defined by inheriting from the Channel class.

Functions

class deepcommpy.channels.Channel(noise_type='awgn')

Channel class for adding noise to input signal

Parameters

noise_typestr

Type of noise to be added to input signal. Valid values are ‘awgn’, ‘fading’, ‘radar’, ‘t-dist’, ‘EPA’, ‘EVA’, ‘ETU’

__init__(noise_type='awgn')

corrupt_signal(input_signal, sigma=1.0, vv=5.0, radar_power=20.0, radar_prob=0.05)

Corrupt input signal with noise

Parameters

input_signaltorch.Tensor

Input signal to be corrupted

sigmafloat

Standard deviation of noise

vvfloat (optional)

Degrees of freedom for t-distribution

radar_powerfloat (optional)

Power of radar noise

radar_probfloat (optional)

Probability of radar noise

Returns

corrupted_signaltorch.Tensor

Corrupted signal