Grey Particle Filter (GPF) for Self-Estimating Depth of Maneuvering Autonomous Underwater Vehicle (AUV)

Abstract

This paper presents a grey particle filter (GPF) that incorporates the grey prediction algorithm into the particle filter (PF). The GPF self-estimates the depth of maneuvering autonomous underwater vehicle (AUV) using the data measured by the depth sensor equipped in the AUV under the condition that the prior maneuvering information is unknown and the measurement noise is time-varying. The principle of the GPF is that the particles are sampled by grey prediction algorithm and the likelihood probabilities of the grey particles are calculated by wavelet transform in real time, which only uses the historical measurement without establishing prior dynamic models. Therefore, the GPF can effectively alleviate the sample degeneracy problem which is common in the multiple model particle filter (MMFP). The performance of the MMPF and GPF are both evaluated through the experimental data. The results show that GPF has the better estimation accuracy than the MMPF.