Quiz collegato alla guida · Medio Livello

Quiz sul filtro di Kalman

Segui i passaggi di previsione e aggiornamento di Kalman, calcola un guadagno scalare e comprendi come l'incertezza controlla la ponderazione della misurazione.

Percorsi guida correlatiFiltro di Kalman
Domanda 1 di 8

Which sequence describes one standard Kalman filtering cycle?

Anteprima di ogni domanda8 domande

Le risposte e le spiegazioni vengono visualizzate mentre rispondi al quiz.

  1. Which sequence describes one standard Kalman filtering cycle?
    • Predict the state, then update it with a measurement
    • Update with a measurement first, then predict the next state
    • Replace the state estimate with each raw measurement
    • Predict the sensor reading but omit the measurement correction
  2. With predicted variance 4 and measurement variance 4 in the scalar case, what is the Kalman gain?
    • 1
    • 0,5
    • 4
    • 8
  3. Using gain 0.5, predicted state 10 and measurement 14, what updated state results?
    • 10
    • 14
    • 12
    • 7
  4. If measurement noise variance increases while prediction uncertainty stays fixed, what happens to the gain?
    • It becomes the measurement variance
    • It increases above one
    • It remains exactly unchanged
    • It decreases, so the measurement receives less weight
  5. Which conditions support the classical Kalman filter's optimality result?
    • A nonlinear process with arbitrary noise distributions
    • Linear dynamics and Gaussian noise with a correctly specified model
    • A stationary mean with no state or measurement covariance model
    • A discrete-state process with arbitrary transition rules
  6. Which residual is called the innovation during measurement update?
    • The accumulated process-noise estimate
    • The covariance of the previous state
    • Observed measurement minus predicted measurement
    • Predicted measurement minus observed measurement
  7. Persistent innovation bias may point to what?
    • A proof that the system is stationary
    • A guarantee that all states are correct
    • A need to set measurement noise to zero
    • A biased sensor or misspecified process model
  8. What does the covariance P represent in the filter?
    • Uncertainty in the estimated hidden state
    • The measured state itself
    • The sensor identity label
    • The number of filter updates