The Particle Filter Algorithm Based on Improved Resampling Techonology

There exists some defects in Particle Filter algorithm based on the traditional resampling techonology such as large calculation, severe particle impoverishment, low estimation accuracy and so on. To solve these problems, this paper proposes an improved Particle Filter algorithm based on resampling techonology. First, the initial particles need to be processed before resampling to get a new set of particles. Then all the particles are divided into two types: one is medium weight particle collection, the other one is large weight and small weight particle collection. The medium weight particle collection remain unchanged without resamping. However, the other one collection first has to be judged whether it can satisfy the condition for resampling. Finally, a linear combination of large weight and small weight particles will be used to get new particles if it meets the resampling requisites.Results of simulation shows that the proposed improved Particle Filter algorithm is feasible.