Introduction

The codes are for comparing the estimation performance of
L1 minimization without prior,
weighted l1 minimization with prior,
weighted l1 minimization with prior Pruning

The theorem foundation can be found in our submitted paper:
RESILIENT STATE RECOVERY USING MEASUREMENT SUPPORT1 PRIOR

If you use the whole codes or part of them, please cite the paper:

@inproceedings{zheng2021attack,
  title={Attack-Resilient Weighted $$\backslash$ell\_ $\{$1$\}$ $ Observer with Prior Pruning},
  author={Zheng, Yu and Anubi, Olugbenga Moses},
  booktitle={2021 American Control Conference (ACC)},
  pages={340--345},
  year={2021},
  organization={IEEE}
}

Any question about the paper, please contact Yu Zheng (yz19b@fsu.edu), Dr.Anubi (anubi@fsu.edu).

Simulation results

Entrance:

total_simulation.m

post-processing entrance

Visulization.m

generate the figures in paper

Main functions:

resilient_estimation_Withoutprior.m

L1-minimization without prior

resilient_estimation_Withoracle.m

Weighted L1-minimization with oracle

resilient_estimation_Withpruning.m

Weighted L1-minimization with pruning algorithm based on oracle

Support functions:

(4) L1_noprior_solver.m (original l1-minimization solver)
(5) Weighted_L1_solver.m (weighted l1-minimization solver)
(6) sysGen.m (generate full observable (A,C))
(7) gen_attack_channel.m (FDIA generator)
(8) pruning_algorithm.m (pruning algorithm)

Small support functions:

(9) CDF_inv.m (generate agreement probability satisfying ROC)
(10) pmf_PB.m (calculate pmf of Poisson-binomial distribution)
(11) reliable_num_attacks.m (for first step in pruning algorithm)