Factors affecting Kalman filter performance

Parent: 1D Kalman filters Source: rlabbe Kalman/Bayesian filters in Python

Difficulties of creating a well-performing Kalman filter: Includes modeling the sensor performance (what variance most accurately represents the reality? Which probability distribution?)

Factors affecting the performance of the Kalman filter

• On modelling the process noise/variance
• Bad initial estimate
  • Filter can recover from this, because we have a certain belief in the sensor measurements
  • Typically the initial value is set to the first sensor measurement
• Nonlinearity of the system