Wavelet Analysis Based Hidden Markov Model for Large Ship Trajectory Prediction

Abstract

Listen

Translate article

For the trajectory prediction problem of large ship, it is difficult to establish accurate motion model of moving target. Traditional trajectory prediction algorithm based on Kalman filtering or linear regression have defects of lower precision or long time delay. In this paper, we discuss the trajectory prediction problem based on Hidden Markov Model (HMM). In order to reduce the influence of error accumulation on prediction accuracy, we add wavelet analysis into HMM framework and propose a wavelet-based HMM trajectory prediction algorithm (HMM-WA). Specifically by wavelet transform and single reconstruction, we convert trajectory sequence into a column vector and then use this vector as input of the HMM. The simulation results show that the HMM-WA algorithm can effectively improve the prediction accuracy compared with the classical HMM, linear regression methods and the Kalman filter.