🗒️Kalman Filter
2025-12-5
| 2025-12-5
Words 341Read Time 1 min
🤖
Formulation and implementation of Kalman filter

Linear Dynamical System

State: is the true status of the system. Measurement: is the measured status of the system.
is called transition matrix. is called observation matrix.
e.g. an agent moving on a 2D plane. ,

KF steps

Initialize

At , we assume the state to be normal distribution.

Predict

Before we got the next measurement, we “guess” the next state distribution.
Posterior from previous steps:
Predictions for new step:
The mean and variance are computed by:

Update

When new measurement arrives, we “correct” the predicted state distribution.
We treat the prediction as prior, then incorporate measurement to get posterior .
The parameters are computed by:
 
  • 技术分享
  • Tutorial 4彭州
    Loading...