An ARIMA-Based Autonomous Underwater Vehicle Tracking Algorithm

Abstract

In order to track the Autonomous Underwater Vehicles (AUVs) in highly noisy underwater environments, we propose an AutoRegressive Integrated Moving Average-based AUV Tracking Algorithm, coined as ARIMA-AT. First, we propose an atomic norm minimization-based beamforming approach to enhance the prediction accuracy by addressing the limitations posed by sparse and incomplete sensor data. Second, we construct a beamforming model to predict the AUV trajectory based on the ARIMA framework, which minimizes the parameter error using a Lagrange multiplier method. This model accounts for temporal dependencies in the AUV’s movement and improves the robustness of tracking in dynamic underwater environments. Third, we propose a Fisher information matrix based parameter prediction method to predict ARIMA-AT parameters for further refining the accuracy of the ARIMA parameter estimation and reducing the impact of noise interference. Experimental results demonstrate that the ARIMA-AT algorithm can reduce prediction errors and accurately track the movement of AUVs in low signal-to-noise ratio (SNR) underwater environments.