Improved Particle Filter Using Clustering Similarity of the State Trajectory with Application to Nonlinear Estimation: Theory, Modeling, and Applications

Algorithm 2

Clustering similarity particle filter.

Step 1. Particle initialization

At time , for

Sample

At time , for

Step 2. Importance sampling

Predict , sample

Step 3. Similarity measurement

Particle state trajectories

 Draw

      Calculate

 Exponential transformation

Step 4. Recursive importance weights

Revised proposal distribution

Update weights

    Normalize

Step 5. State estimation