Safety-critical control is crucial but difficult in the applications of autonomous underwater vehicles (AUVs). This article proposes a novel hierarchical safety-critical framework for the control of AUVs consisting of waypoint-based optimal trajectory generation and tracking. The objective is to generate a smooth and feasible trajectory to conform to the dynamics of the AUV and then control the AUV to operate in a safe region. To this end, the offline trajectory generator and the online controller are designed. In the offline step, an improved minimum snap polynomial trajectory is generated as the tracking reference. Specifically, the trajectory is optimized by explicitly considering the practical constraints of the AUV's actuator, which alleviates the complexity of the controller design and computational burden compared with the real-time nonlinear optimization. In the online step, the time- and state-dependent high-order control barrier functions are incorporated into optimization through quadratic programming (QP) that modifies the backstepping-based nominal controller in a minimally invasive way. As a supplement and adjustment to the offline planning, the online step ensures that the system states are maintained in the safety set, thus ensuring obstacle avoidance. The online computational efficient QP structure guarantees convenience and scalability in practical implementation. Both the online static and dynamic obstacle avoidance simulations demonstrate the adherence to the safety constraints. Experimental results validate the effectiveness of the proposed method.

This article proposes a robust safety–critical control framework based on high order control barrier function (HOCBF) and disturbance observer (DOB) for control affine systems in the presence of unknown disturbances. Specifically, we apply the proposed framework for the obstacle avoidance of autonomous underwater vehicles (AUVs) to consider the nonlinear characteristics and current disturbance. The system stability is proved to be uniformly ultimately bounded (UUB). The proposed framework first estimates the unknown term of the Lie derivative of HOCBF by employing a first-order high-gain DOB. The bounded property of the DOB estimation error is combined with the Tunable Input-to-State Safe (TISSf) to compensate for the forward invariant condition of CBF. We construct the DOB-TISSf-based robust safety–critical control framework based on quadratic programming to guarantee safety and control performance simultaneously. We show that the safety–critical framework still renders the system UUB. Simulations and field experiments are validated on AUV to avoid obstacles. The results demonstrate that the proposed framework ensures the safety of AUV. Moreover, the framework allows for lower conservativeness of the disturbed system and thus a better tradeoff between system performance and safety.

This letter focuses on the constraint-driven optimal control for multiple Autonomous Underwater Vehicles (AUVs). The different tasks are formulated as multiple constraints on optimization. First, the task goals are formulated by the time-varying control barrier functions. Then, a novel cost-to-constraint transformation method is proposed for time-varying tasks to achieve the high-order constraint-driven task execution under AUV dynamics. The priority of each task is adjusted by the maximum allowed limit of the slack variable, thus relaxing the task constraints and avoiding potential constraint conflicts. Finally, the collision-free optimal energy-aware task execution is achieved by a minimum input norm objective that is incorporated into computationally efficient quadratic programming, which is scalable to different task behaviors. Comparative numerical simulations considering time-varying trajectory tracking and cooperative coverage tasks validate the effectiveness of the proposed method.

The underwater leveling motion control of a morphable unmanned aerial–aquatic vehicle (M-UAAV) is essential but lacks an elegant solution. In this article, a scheme named Aqua Slide is proposed for Mirs-Alioth, our predeveloped M-UAAV prototype, to achieve underwater leveling motions by utilizing singularity. For the first time, the singular thrust tilt angle (STTA), i.e., the tilt angle that causes a vehicle to reach singularity, is characterized and defined. For underwater leveling motion generation, STTA is utilized as a key tool. Crucial factors of these motions, i.e., the existence of control direction uncertainty, and coupling effects are identified. Aqua Slide is proposed to tackle the crucial factors in generating underwater leveling motions, consisting of an STTA controller, a primary-auxiliary switch, and an auxiliary controller incorporating the saturated Nussbaum function. A simulation environment, enriched with a detailed model, is established using a Gazebo and hardware-in-the-loop (HITL) configuration. Under similar control parameter settings, full-cycle Gazebo-HITL simulations and experiments reveal the effectiveness of the approach. The vehicle maintains the attitude within a maximum of 5${}^{\circ}$ during the whole procedure. Meanwhile, the ablation studies in Aqua Slide highlight the indispensability of the submodules in the approach. The absence of any module results in undesirable or even divergent behavior of Mirs-Alioth during underwater leveling motions. The preliminary exploration in this article elegantly solves the issue of generating underwater leveling motion for an M-UAAV and provides a referential solution for transitioning from simulation to real-world applications. Video attachment: https://youtu.be/XpViXL0qcvk.

Reinforcement learning (RL) has been shown to be effective in path planning. However, it usually requires exploring a sufficient number of state–action pairs, some of which may be unsafe when deployed in practical obstacle environments. To this end, this paper proposes an end-to-end planning method based model-free RL framework with optimization, which can achieve better learning performance with a safety guarantee. Firstly, for second-order drone systems, a differentiable high-order control barrier function (HOCBF) is introduced to ensure the output of the planning algorithm falls in a safe range. Then, a safety layer based on the HOCBF is proposed, which projects RL actions into a feasible solution set to guarantee safe exploration. Finally, we conducted a simulation for drone obstacle avoidance and validated the proposed method in the simulation environment. The experimental results demonstrate a significant enhancement over the baseline approach. Specifically, the proposed method achieved a substantial reduction in the average cumulative number of collisions per drone during training compared to the baseline. Additionally, in the testing phase, the proposed method realized a 43% improvement in the task success rate relative to the MADDPG.

In the collaborative application of multiple autonomous underwater vehicles (AUVs), collision avoidance is essential for ensuring the physical safety of AUVs, while a specific connectivity topology is required for information exchange. This dual requirement presents challenges for controller synthesis. Thus, this study focuses on collision-free connectivity topology switching and maintenance during collaborative missions. For the primary control goal of reaching predefined locations, a backstepping-based controller is designed to ensure system stability. Collision avoidance is achieved through the application of high-order zeroing control barrier functions (ZCBFs), which guarantee the forward invariance of the desired safety set. To facilitate high-bandwidth information exchange, establishing a specific connectivity topology within an effective range is necessary. High-order finite-time convergence control barrier functions (FCBFs) are employed to form new topologies within finite time and then maintain the forward invariance of the desired set. The controller is synthesized using quadratic programming optimization, incorporating collision-free topological constraints and minimally modifying the backstepping-based controller. Numerical simulations and field experimental results demonstrate the effectiveness of the proposed controller, showing that topology switches faster than in existing studies while ensuring collision avoidance under disturbances.