Glocal trajectory generation and tracking control for autonomous underwater vehicles with optimal coverage sensor networks

This paper presents a comprehensive framework for glocal trajectory generation with real-time tracking control for a group of Autonomous Underwater Vehicles (AUVs) equipped with distributed sensors. A two-stage approach is proposed to maximize the underwater area coverage of sensor systems while ensuring network connectivity between AUVs and free collision with terrains and floating obstacles. At the global level, a heuristic algorithm named Global Trajectory to Maximize Coverage (GT-MC) is introduced, which generate trajectory to optimize the final AUVs distribution. After that, the trajectory is further optimized to produce the final set of waypoints for the AUVs group. At the local level, a safety-critical trajectory generation method is developed by using a Model Predictive Control (MPC) scheme for a virtual AUV system with Control Barrier Functions (CBF) as constraints for floating obstacle avoidance. Then, the generated trajectories are tracked by the actual AUVs using a base controller, in this case a classical Sliding Mode Controller (SMC) combined with a thruster force allocation optimizer. The complete framework is validated via simulation studies using an open-source advanced physics tool called MuJoCo. The suggested methodology can facilitate the autonomy, scalability, and safety of sensor-AUVs distribution missions, making it a promising tool for intelligent marine sensing and monitoring.