Autocovariance Least Squares with Constrained Noise Covariance Model Identification
The scripts to run the constrained Autocovariance Least Squares technique for two examples are contained in this repository.
-
gen_data_msd
- Creates simulated datasets of the msd dynmaics with process and measurement noise
- Generates ALS inputs with initial suboptimal process noise covaraince,
$Q$ and measurement noise covariance,$R$ .
-
run_als_msd
- Run script for constrained ALS problem
- Calls setup_ALS_msd.m with defines lags and other ALS inputs
-
als_msd
- ALS class with mass spring damper constraints for 7 temperature problem
-
plot_lags_msd
- Plots results from constrained ALS problem
- Saves mean and standard deviation of
$Q$ and$R$ solutions - Figure 2 and Figure 3
-
plot_QR_T
- Figure 5
-
wtData_setup
- Loads wind tunnel datasets: https://github.com/uwaa-ndcl/ACC_2025_Hinson/tree/main/MARGE/Data
- Relies on models locted here: https://github.com/uwaa-ndcl/ACC_2025_Hinson/tree/main/MARGE/Models
- Generates ALS inputs with initial suboptimal process noise covaraince,
$Q$ and measurement noise covariance,$R$ .
-
run_als_MARGE
- Run script for constrained ALS problem
- Calls setup_ALS_MARGE.m with defines lags and other ALS inputs
-
als_MARGE
- ALS class with MARGE constraints for 4 dynamic pressure problem
-
plot_lags_MARGE
- Plots results from constrained ALS problem
- Saves mean and standard deviation of
$Q$ and$R$ solutions
-
plot_QR_qbar
- Figure 7
- Relies on unconstrained ALS solutions:https://github.com/uwaa-ndcl/ACC_2025_Hinson/tree/main/MARGE/Results/Unconstrained
-
wtData_ALSest
- Figures 8-13
- Relies on wind tunnel data: https://github.com/uwaa-ndcl/ACC_2025_Hinson/tree/main/MARGE/Data/WT_post
