wuyou33 / master_thesisekf

Parameter Identification via Kalman Filter on a model motor glider

Master_thesis

Parameter Identification via Kalman Filter on a model motor glider

---

Structure of folders and files

simulink_model

• testEnvironment.slx: Test environment including model of the Vitesse and added sensors (Meyer-Brügel, Modifications by me)
• init_testEnvironment.m: initializes and opens test environment
• Init: contains scripts and data for initialization (Meyer-Brügel)

kalman_filter

• main.m: runs filters (basic structure by Jategaonkar, Modifications by me)
• xdot.m: Equations of motion of the Vitesse
• obs.m: Output Equation
• ruku4.m: numerical integration method, Runge-Kutta 4
• ekf.m: algorithm of the EKF (Jategaonkar)
• ukf.m: algorithm of the UKF with noise estimation (Jategaonkar)
• ukf_mod.m: algorithm of the UKF with additive zero-mean noise (Jategaonkar)
• sysmatA.m: linearizes state matrix (for EKF) (Jategaonkar)
• sysmatC.m: linearizes output matrix (for EKF) (Jategaonkar)
• print_par_std.m: prints average estimated parameter, standard deviation and relative standard deviation (Jategaonkar, Modifications by me)
• plot_params.m: plot results of parameter estimation
• plot_states_outputs.m: plot results of state estimation\ output prediction together with measurements
• mDef_check.m: checks if dimensions are set correctly (Jategaonkar)
• mDef.m: Definition of model, flight data, initial values etc. (Jategaonkar, Modifications by me)

flight_test_eval

• read_logs.m: saves maneuver data for selected flight
• plot_maneuvers.m: plots all maneuver data
• find_std_dev_log.m: computes standard deviations and variances of measured output to use as inital value for optimization of rr and for sensor models in simulink
• optimize_filters.m: inlcudes optimization modes to find suitable maneuvers, system noise covariance Q, scaling parameter alpha and measurement noise cov. R
• logs
  • contains log files
• testflight_maneuver_log_data
  • contains generated maneuver data from read_logs.m