Skip to content

Time-varying frequency Estimation of narrow-band signals via Extended Kalman Filter and Unscented Kalman Filter. The two methods are compared and a thorough study on the influence of the parameters is performed, along with numerical considerations. Assignment for the Model Identification and Adaptive Systems course @POLIMI, 2017-2018

License

Notifications You must be signed in to change notification settings

infinitesnow/EKFUKF

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

40 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Time-varying frequency Estimation of narrow-band signals via Extended Kalman Filter and Unscented Kalman Filter

See included paper for more details.

Files list

  • main_audio.m Performs frequency tracking of 3 audio files at 3 different frequencies. Generates a plot of the estimation.
  • main_pi.m Performs analysis of the Performance Index and generates plots and data in the data/pi folder. Iterates over several values of q and sigma_noise
  • main_ri.m performs analysis of the Robustness Index and generates plots and data in the data/ri folder.

Flags

main_pi.m

  • LOGSPACE determines if the array of sigmas is spaced linearly (false) or logarithmically (true)

ekf/ekf.m

  • SAVE_EKF_PLOT if true, generates a video in generated_video/ of the estimation process.
  • LOG if true, generates a complete log of the values during the estimation
  • VERBOSE if true, prints to console the iteration number. Useful in conjunction with SAVE_EKF_PLOT to check on the progress, which makes the process really slow.

ukf/ukf.m

  • VERBOSE if true, prints to console the iteration number. Useful in conjunction with SAVE_EKF_PLOT to check on the progress, which makes the process really slow.
  • LOG if true, generates a complete log of the values during the estimation
  • SAVE_UKF_PLOT if true, generates a video in generated_video/ of the estimation process.
  • SAVE_SP_PLOT if true, generates a video in generated_video/ showing the motion of sigma points in the state space.

pi/pi_plot.m

  • PLOT_FIT if false, plots raw data cloud. If true, plots the Least Square fit of a hyperbola of the data

main_ri.m

  • PLOT_PREDICTION if true, generates a video in generated_video/ with an iteration for every frame
  • PLOT_PROFILE if true, opens a plot with the frequency profile of the generated signal
  • COMPUTE if false, just plots without computing, loading previously generated data.

About

Time-varying frequency Estimation of narrow-band signals via Extended Kalman Filter and Unscented Kalman Filter. The two methods are compared and a thorough study on the influence of the parameters is performed, along with numerical considerations. Assignment for the Model Identification and Adaptive Systems course @POLIMI, 2017-2018

Resources

License

Stars

Watchers

Forks

Packages

No packages published