An AUV-TTN Scheme Applied to AUV Trajectory Tracking Under Noise Disturbance

In marine engineering, autonomous underwater vehicle (AUV), combining self-navigating, decision-making, and operational skills, is crucial for oceanic missions. And trajectory tracking is a key metric for assessing AUV performance and has received considerable focus. However, existing AUV trajectory tracking methods exhibit limitations in control accuracy, and demands substantial computational resources and time, thereby constraining their real-time applicability in dynamic environments. To tackle these challenges, this paper simplifies the AUV trajectory tracking problem into an inequality and equality constrained dynamic quadratic programming (IE-DQP) problem and introduces the zeroing neural network (ZNN) model to address it. Additionally, to reduce underwater noise disturbances, we propose the projection and elliptical ZNN (PEZNN) model, utilizing the Frobenius norm, and develop two projection functions that align with the physical constraints of AUVs. Collectively, we present the aforementioned approach as the AUV-TNN scheme. To assess its performance, we conduct rigorous mathematical derivations to demonstrate the convergence and robustness of the proposed scheme. Finally, the effectiveness of the proposed approach is confirmed through three comprehensive simulated AUV trajectory tracking experiments.