Particle Filter for Positioning Accuracy Improvement in GNSS Receiver

Abstract

Traditional GNSS receiver usually uses Least Squares (LS) method or Kalman filtering to calculate the position. This paper proposes a new method that uses particle filtering for positioning calculation. The proposed particle filtering bases on LS method, but resolves the low positioning accuracy problem of LS. It also overcomes the Kalman filtering's drawback that need to know noise properties and receiver's dynamic model previously. The paper gives detailed implementation of the particle filter that is applicable to positioning calculation in GNSS receiver. Based on a suitably built model, the filter uses a modified resampling algorithm, and selects reasonable particle number and proposal distribution according to simulation test. These effectively reduce the calculation and improve the performance of the filter. Simulation results with GSS6700 simulator, based on the digital Intermediate Frequency (IF) signals, show that compared with traditional LS method and Kalman filtering, the proposed particle filtering achieves much higher positioning accuracy.