Constrained Path Following Control of AUVs with Model Predictive Guidance: A Periodic Dynamic Event-Triggered Approach

This paper presents a Lyapunov-based model predictive guidance (LBMPG) approach for constrained path-following control of underactuated autonomous underwater vehicles (AUVs). The proposed method generates optimal guidance signals, subject to velocity and heading error constraints, using an online Lyapunov-based model predictive control (MPC) framework with a stability-enforcing contractive constraint. To reduce the computational burden of frequent optimization, a novel periodic dynamic event-triggered mechanism (PDETM) is introduced, which activates optimization based on a triggering condition tied to prediction accuracy and stability preservation. The resulting closed-loop system, characterized by mixed continuous and discrete dynamics, is analyzed within a hybrid system framework. Sufficient conditions are derived to ensure stability by jointly constraining the event-detection period, triggering functions, and related parameters. Simulations validate that the approach has effective path-following performance with significantly reduced computational demands, aligning with practical AUV requirements.