Channel Model Blocks

Introduction

This is the gr-channels package. It contains signal processing blocks to simulate channel models.

The Python namespace is in gnuradio.channels, which would be normally imported as:

from gnuradio import channels

See the Doxygen documentation for details about the blocks available in this package. A quick listing of the details can be found in Python after importing by using:

help(channels)

Available Channel Models

AWGN Channel

The basic gr::channels::channel_model block is essentially an additive white Gaussian noise (AWGN) channel with a few extra additions. This channel model simulates AWGN as well as frequency and timing offsets between the transmit and receiver and a simple static multipath environment.

The parameters include:

• noise_voltage The AWGN noise level as a voltage (to be calculated externally to meet, say, a desired SNR).
• frequency_offset The normalized frequency offset. 0 is no offset; 0.25 would be, for a digital modem, one quarter of the symbol rate.
• epsilon The sample timing offset to emulate the different rates between the sample clocks of the transmitter and receiver. 1.0 is no difference.
• taps Taps of a FIR filter to emulate a multipath delay profile.
• noise_seed A random number generator seed for the noise source.

Fading Channel Model

The gr::channels::fading_model builds a basic fading simulation. This model approximates a fading channel using the sum of sinusoids method for the number of expected multipath components. This block also takes in the Doppler frequency shift as a normalized value, a line-of-sight (LOS) parameter that is either true or false, the Rician K factor, and a random seed to the noise generators. These parameters are provided in the documentation as:

• N the number of sinusoids to use in simulating the channel; 8 is a good value
• fDTs normalized maximum Doppler frequency, fD * Ts
• LOS include Line-of-Site path? selects between Rayleigh (NLOS) and Rician (LOS) models
• K Rician factor (ratio of the specular power to the scattered power)
• seed a random number to seed the noise generators

Frequency-Selective Fading Model

Extending the fading model, the gr::channels::selective_fading_model channel model allows us to specify frequency selective parameters for the model. Starting with the same inputs as the gr::channels::fading_model, this block also takes in power delay profile (PDP). The PDP is given as a vector of the times of the delays and another vector of the corresponding magnitudes of the PDP. We also specify the number of taps in the filter to simulate this. These parameters are provided in the documentation as:

• N the number of sinusoids to use in simulating the channel; 8 is a good value
• fDTs normalized maximum Doppler frequency, fD * Ts
• LOS include Line-of-Site path? selects between Rayleigh (NLOS) and Rician (LOS) models
• K Rician factor (ratio of the specular power to the scattered power)
• seed a random number to seed the noise generators
• delays a vector of values that specify the time delay of each impulse
• mags a vector of values that specify the magnitude of each impulse
• ntaps the number of filter taps.

Hardware Impairments Model

GNU Radio also provides a model for incorporating parameters of a radio transmitter and receiver to better simulate signals as they would be transmitted and received through real hardware. This simulation tool can be used to see how a waveform would behave over a known radio, or if designing radio hardware, this block could help understand the requirements and tolerances for the new system.

The hardware impairments model is a Python-only hierarchical block, which is declared as digital.impairments(...). The block is also represented in GRC under the "Impairment Models" category.

The impairments model is actually made up of a handful of other hierarchical blocks, each of which perform one aspect of the impairments model. Each subsection is usable on its own, and each one can be used in GRC. These subsystems are:

• digital.phase_noise_gen ("Phase Noise Generator")
• digital.iqbal_gen ("IQ Imbalance Generator")
• digital.distortion_3_gen ("Third Order Distortion")
• digital.distortion_2_gen ("Second Order Distortion")

There are also a few other blocks that were built to develop and simulate different hardware behaviors

• digital.quantizer ("Quantizer"): quantizes to b number of bits.
• digital.amp_bal ("Amplitude Balance"): corrects IQ amplitude imbalance.
• digital.phase_bal ("Phase Balance"): corrects IQ phase imbalance.

The digital.impairments model combines the phase noise, IQ imbalance, and second and third order distortion generators. It also performs I and Q offset imbalance and internal frequency shifting problems.

The parameters for using the hardware impairments model are:

• phase_noise_mag: the level of the phase noise (in dB).
• magbal: The level of the magnitude imbalance between I and Q.
• phasebal: The level of the phase imbalance (as an angle).
• q_ofs: Offset value of the quadrature arm (as a magnitude).
• i_ofs: Offset value of the in-phase arm (as a magnitude).
• freq_offset: normalized frequency offset.
• gamma: Level of the 2nd order distortion.
• beta: level of the 3rd order distortion.