This is a communication system simulation code for nonlinearly distorted OFDM with receivers based on generalized approximate message passing algorithm. For more details, see Ref [1, 2].
This simulation code relies on Armadillo (C++ library for linear algebra & scientific computing). Other than that there are no external dependencies.
Firstly, you need to install Armadillo C++ library. You can download a copy of Armadillo C++ library from the official web-site (http://arma.sourceforge.net/), or you can use pre-built Armadillo packages provided by many Linux-based operating systems. Personally, I use Visual Studio 2019 and install Armadillo via vcpkg package manager (https://github.com/Microsoft/vcpkg).
The principal way to run simulation is via command line interface with simulation parameters specified in cfg.txt file.
The parameters that can be specified in cfg.txt file are given in the table below:
| Parameter | Description |
|---|---|
| N | the number of subcarriers in oversampled signal |
| Nu | the number of nonzero subcarriers (i.e., oversampling factor M=N/Nu) |
| I_max | the maximum number of iterations |
| G0 | the input signal scaling factor |
| s_scale | the noise variance scaling factor |
| B | the damping factor |
| rand_seed | the random seed |
| EbNo | the signal-to-noise ratio per bit (array, e.g. 7.75 8.0 8.25) |
| L_SNR | the number of OFDM symbols to simulate per EbNo (array) |
| qam_order | the QAM modulation order (supported: 4, 16, 64) |
| enable_multipath_channel | set this value to 1 to simulate Rayleigh block-fading channel |
| Channel taps | the number of channel taps |
| Channel delay spread | the channel RMS delay-spread (normalized) |
| enable_oob_filtering_tx | set this value to 1 to enable out-of-band filtering in the transmitter |
| enable_oob_filtering_rx | set this value to 1 to enable out-of-band filtering in the receiver |
| post_correction | set this value to 1 to enable 1-bit correction post-processing |
N = 16384
Nu = 4096
I_max = 75
G0 = 0.0
s_scale = 0.71
B = 0.875
rand_seed = 12345
EbNo = 7.75 8.0 8.25
L_SNR = 1000 1000 1000
qam_order = 16
enable_multipath_channel = 0
enable_oob_filtering_tx = 0
enable_oob_filtering_rx = 0
post_correction = 1
The simulation results are stored in:
- EbNo.txt (all simulated EbNo values)
- BER.txt (the BER result for each EbNo value)
- FER.txt (the FER result for each EbNo value)
- log.txt (all the output you see on the screen during simulation)
- S. Zhidkov and R.Dinis “Belief Propagation Receivers for Near-Optimal Detection of Nonlinearly Distorted OFDM Signals,” 2019 IEEE 89th Vehicular Technology Conference (VTC2019-Spring), May 2019 Download
- S. Zhidkov, "Orthogonal transform multiplexing with memoryless nonlinearity: a possible alternative to traditional coded-modulation schemes,“ in Proc. 9th International Congress on Ultra Modern Telecommunications and Control Systems (ICUMT-2017), Munich, Germany, November, 2017 Download
- Sergey Zhidkov - Initial work (https://cifrasoft.com/people/szhidkov/)
This project is licensed under the MIT License - see the LICENSE.md file for details