A Survey on Sensor-based Planning and Control for Unmanned Underwater Vehicles

A team of researchers including Shivam Vishwakarma and Tejal Bedmutha has published a comprehensive survey on arXiv (paper ID: 2604.05003) examining sensor-based planning and control methods for Unmanned Underwater Vehicles (UUVs). The work addresses the unique challenges of underwater navigation where vehicles operate without Global Navigation Satellite System (GNSS) signals, contend with drifting currents and noisy sensor data, and face low-bandwidth, high-latency acoustic communications. The survey's core focus is on reactive local planning layers that process real-time inputs from sensors like SONAR and Inertial Measurement Units (IMU) to improve localization and enable autonomy in dynamic ocean environments.

The researchers provide a taxonomy that categorizes existing approaches into two main architectures: decoupled and coupled systems. Decoupled architectures handle planning and control sequentially, while coupled architectures create tighter feedback loops for more immediate responsiveness to environmental changes. The comparative analysis reveals trade-offs between different control methodologies: PID controllers offer simplicity but lack predictive capability for complex maneuvers; Model Predictive Control (MPC) provides superior path optimization at higher computational cost; and invariant-set controllers deliver strong safety guarantees but may sacrifice agility in confined spaces.

Key contributions include developing a structured framework for understanding how planning and control integrate in UUV systems, with particular emphasis on adaptive local planning strategies. The survey analyzes how different controllers function within integrated planning frameworks, providing insights for researchers and engineers developing next-generation autonomous underwater systems capable of dynamic obstacle avoidance and on-the-fly re-planning in uncertain conditions.