Research on Underwater Visual Docking Control Method for Dual-AUVs

Underwater docking facilitates autonomous underwater vehicles (AUVs) in energy replenishment and data transmission, extending their operational range. This study focuses on a dual-AUV system comprising a target AUV with visual markers and a docking AUV equipped with an optical camera. Addressing challenges such as image delay and loss, which cause delayed control inputs, we first employ the autoregressive integrated moving average (ARIMA) algorithm to predict the relative orientation and position between the AUVs. ARIMA’s robust time-series forecasting capability enhances prediction accuracy, mitigating delay effects and ensuring smoother docking. Subsequently, we design fuzzy rules based on image delay time and loss rate, developing a fuzzy PID control for adaptive parameter adjustment. This approach leverages fuzzy logic to handle uncertainties and nonlinearities in underwater environments, enhancing robustness against disturbances and dynamic variations. Experimental validation in a swimming pool demonstrates that our method improves control input accuracy from 85% to 92% and reduces overshoot from 15% to 5%, significantly enhancing system performance and control precision for dual-AUV visual docking.