Skip to content

Latest commit

 

History

History
44 lines (33 loc) · 2.33 KB

README.md

File metadata and controls

44 lines (33 loc) · 2.33 KB
Raw

DI-S-MMSE Turbo Equalization for OTFS Modulation

paper paper License: MIT

About

This repository provides code to reproduce the results of DI-S-MMSE Turbo equalizer, which iteratively exchanges the extrinsic information between a soft-input-soft-input (SISO) MMSE estimator and a SISO decoder. Simulation results demonstrate that our proposed equalizer not only approaches the BER performance of the near-optimal symbol-wise MAP algorithm, but also has a linear complexity.

Details of this work can be found on IEEE or arKiv.

Citation

If you find this code useful for your research or use it in your work, please consider citing:

@article{li2023doubly,
  title={Doubly-Iterative Sparsified MMSE Turbo Equalization for OTFS Modulation},
  author={Li, Haotian and Yu, Qiyue},
  journal={IEEE Transactions on Communications},
  volume={71},
  number={3},
  pages={1336--1351},
  year={2023},
  publisher={IEEE}
}

Instructions

Our code is developed and tested on MATLAB 2019a.

Simulation Settings

  • DI-S-MMSE equalizer: (i.e., without decoder) execute DI_S_MMSE.m with your preferred parameters.
  • DI-S-MMSE turbo equalizer: run DI_S_MMSE_Turbo.m with your preferred parameters.
  • LMMSE equalizer: execute DI_S_MMSE.m and set iterTimes=1. This is because the first turbo iteration of DI-S-MMSE equalizer is equivalent to the LMMSE equalizer.

Speedup Trick

If your computer has multiple cores, you can utilize the "parfor" loops to execute this code in parallel on workers, which can significantly speed up the calculations. Specifically, you can replace the line "for sim = 1:sumSim" with "parfor sim = 1:sumSim" in both DI_S_MMSE.m and DI_S_MMSE_Turbo.m.

However, please note that when using the "parfor" loop, you will also need to comment out the display code at the end of the loop.

Contact

Should you have any questions or find any errors, please do not hesitate to let me know:{[email protected]}