I am trying to use an EKF to filter/predict a noisy sinusoidal waveform using the states of phase, rate of change of phase and amplitude. I am finding, however, that sometimes the EKF amplitude state estimate is a negative value which, in the context of my problem does not make any sense, although a negative amplitude is mathematically plausible as it can represent a phase shift that flips or inverts the waveform.

Is there any way to limit the state evaluations within the EKF such that the amplitude will always be a positive value? This will, of course, also have implications for the estimation of the phase state?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.