cs.ROApr 6, 2026
SaveA Survey on Sensor-based Planning and Control for Unmanned Underwater Vehicles
Abstract
This survey examines recent sensor-based planning and control methods for Unmanned Underwater Vehicles (UUVs). In complex, uncertain underwater environments, UUVs require advanced planning and control strategies for effective navigation. These vehicles face significant challenges including drifting and noisy sensor measurements, absence of Global Navigation Satellite System (GNSS) signals, and low-bandwidth, high-latency underwater acoustic communications. The focus is on reactive local planning layers that adapt to real-time sensor inputs such as SONAR and Inertial Measurement Units (IMU) to improve localization accuracy and autonomy in dynamic ocean conditions, enabling dynamic obstacle avoidance and on-the-fly re-planning. The survey categorizes the existing literature into decoupled and coupled architectures for sensor-based planning and control. The decoupled architecture sequentially addresses planning and control stages, whereas coupled architectures offer tighter feedback loops for more immediate responsiveness. A comparative analysis of coupled planning and control methods reveals that while PID controllers are simple, they lack predictive capability for complex maneuvers. Model Predictive Control (MPC) offers superior path optimization but can be computationally intensive, and invariant-set controllers provide strong safety guarantees at the potential cost of agility in confined environments. Key contributions include a taxonomy of architectures combining planning and control, a focus on adaptive local planning, and an analysis of controller roles in integrated planning frameworks for autonomous navigation of UUVs.
Figures & tables
| Application Domain | Description | Corresponding academic reference |
|---|---|---|
| Deep-Sea Exploration | Energy-conserving, collision-free path planning in unknown marine environments | ( [ 4 ] ), ( [ 5 ] ) |
| Environmental Monitoring | Continuously tracking environmental conditions to assess and protect global ecosystems | ( [ 6 ] ), ( [ 7 ] ) |
| Military Operations | Securing seas, projecting power, defending national interests | ( [ 8 ] ), ( [ 9 ] ) |
| Underwater Inspection | Underwater drones navigate using diverse sensors for extended, pilot-less missions | ( [ 10 ] ), ( [ 11 ] ) |
| Scientific Research | Advanced tech for complex underwater exploration and mapping | ( [ 12 ] ), ( [ 13 ] ) |
TABLE I: Applications of UUVs:
Fig. 1: UUV Navigation Stack Hierarchy
Fig. 2: Path Planning Techniques
Fig. 3: A chronological view of Major Technological Breakthroughs in Sensor-based Autonomous Underwater Navigation, its evolution from the 1970s to the 2025.
| Planner/Control Method | Significance | Limitations | Year | Referances |
|---|---|---|---|---|
| Dubins Path + PID | Provides a lightweight and easy-to-implement solution for fixed-curvature path tracking in known environments. Widely used in structured scenarios such as pipeline inspection or area coverage. | Incapable of responding to sudden environmental changes or dynamic obstacles. Performance drops significantly in turbulent or unknown environments. Due, to limited disturbance rejection capability | 2007 | [ 68 ] |
| Deep RL-based Planning | Enables learning of control policies directly from environmental interactions, allowing adaptability in highly dynamic and uncertain underwater scenarios. | Demands large-scale data and extensive training, which may be infeasible in remote deployments. Computational load hinders real-time use on embedded systems. | 2022 | [ 11 ] |
| CBF + MPC | Combines constraint enforcement via Control Barrier Function (CBF) with predictive optimisation via Model Predictive Control (MPC), enabling real-time decision-making that ensures safety in proximity to obstacles and variable currents. | Requires high computational resources, limiting deployment on low-power embedded systems. Real-time performance depends on solver efficiency. | 2020 | [ 33 ] |
| Learning-based Terrain-Aided Navigation | Leverages terrain features and deep models to improve localisation in GNSS-denied environments. Suitable for long-term missions in topographically rich underwater regions. | Relies on accurate prior terrain maps or effective feature extraction. Performance questionable in flat or textureless seafloor regions. | 2023 | [ 48 ] |
| Integrated Planning and Controls (IPC) | Facilitates the joint design of planning and control, minimising trajectory-tracking errors through an integrated model. Enhances robustness in multi-disturbance scenarios. | Optimal performance is questionable. | 2024 | [ 43 , 44 ] |
TABLE II: Widely used Planning and Control Methods in UUVs
| Technology | Direction | Advantages | Disadvantages |
|---|---|---|---|
| Optical | Local directional | Very high data rates (up to hundreds of Mbps), Low-cost implementation. | Highly sensitive to turbidity and particle presence, Requires precise alignment, Very limited range. |
| Radio | Omni-directional | Can penetrate ice and turbid water, Immune to multipath effects, Offers high bandwidth (from several kbps to Mbps). | Easily disrupted by electromagnetic interference, Significantly limited range underwater. |
| Electro-communication | Omni-directional | Compact in size, Energy efficient, Moderate bandwidth, Cost-effective. | Limited transmission distance. |
| Acoustic | Omni-directional | Long operational range (up to 20 km), Proven and mature technology. | Lower bandwidth, Subject to latency and ambient noise. |
TABLE III: Comparison of Underwater Communication Technologies
| Localization Methods | Advantages | Disadvantages |
|---|---|---|
| Dead Reckoning | Autonomous operation with consistent tracking and decent short-term precision | Prone to cumulative drift over time, leading to large positioning errors in extended missions |
| Signal-Based localization Techniques (e.g., Sonar, LBL) | High localization accuracy and immunity to drift enable precise positioning without cumulative error | Susceptible to signal degradation, interference and obstruction; requires infrastructure and has limited range |
| Map-Matching Navigation | Drift-free operation and independence from external signals | Heavily reliant on the availability of accurate and recent maps; high processing demand; performance varies with environmental complexity |
TABLE IV: Comparison of UUVs localisation Techniques
Fig. 4: Elements of UUV navigation stack
Fig. 5: Illustrative model outlining the typical UUV operational setting in an underwater setting
Fig. 6: Illustrative model outlining the typical Decoupled Architecture
Fig. 7: Illustrative model outlining the typical Coupled Architecture
Fig. 8: Scenario 1 : Underwater Environment with non-uniformly distributed unknown static obstacles
Fig. 9: Scenario 2 : Underwater Environment with a static narrow passage
Fig. 10: Scenario 3 : Underwater Environment with a infinitely wide static column ahead of the vehicle.
| Scenarios | Controller | Planner and Controller Metrics | |||||
|---|---|---|---|---|---|---|---|
| Tracked Path Length(m) | Time Taken(s) | Avg Accel. | Max Accel. | Avg Curv. | Max Curv. | ||
| Cluttered Obstacles | MPC | 155.82 | 180.82 | 0.204 | 3.75 | 0.027 | 0.873 |
| PID | 171.12 | 270.46 | 0.226 | 3.46 | 0.069 | 3.800 | |
| Invariant-set | 140.04 | 162.74 | 0.131 | 3.00 | 0.055 | 2.920 | |
| Narrow Passage | MPC | 156.67 | 189.87 | 0.053 | 0.846 | 0.080 | 1.970 |
| PID | 172.77 | 276.49 | 0.105 | 3.90 | 0.030 | 2.980 | |
TABLE V: Performance Metrics Across Navigation Scenarios
| Controller | Controller Metrics | |||||
|---|---|---|---|---|---|---|
| Time Taken(s) | Obstacle Proximity at Crossing (m) | Avg Accel. | Max Accel. | Avg Curv. | Max Curv. | |
| MPC | - | - | - | - | - | - |
| PID | 724.6 | 25.75 | 0.01 | 2.81 | 0.049 | 0.19 |
| Invariant-set | 1054 | 3.81 | 0.025 | 0.43 | 0.015 | 0.57 |
TABLE VI: Performance Metrics of controllers on Figure-Of-Eight shape curve
Fig. 11: Scenario 4 : Underwater Environment with a infinitely long static column ahead of the vehicle.
Fig. 12: Scenario 5 : Underwater Environment with Temporally Varying Dynamic Obstacles in the Vehicle’s Field of View .
Fig. 13: Tracked Path visualization of Figure-Of-Eight curve using Invariant-set controller
Fig. 14: Tracked Path visualization of Figure-Of-Eight curve using PID controller
| Controller | Key Advantages | Key Limitations | Suitable Scenarios | Less Suitable Scenarios |
|---|---|---|---|---|
| PID | • Simplicity and ease of implementation. • Effective in linear environments with minimal oscillations. | • Lacks predictive capabilities; struggles with sudden changes. • Higher acceleration peaks and sharp turns. • No inherent constraint handling or obstacle avoidance. | • Linear path following. • Simple waypoint tracking. | • Dynamic obstacle avoidance. • Complex maneuvering. • Highly constrained environments. |
| MPC | • Time-effective performance and path optimization. • Reduces deviations and maintains goal bias. • Strong performance in narrow passages by balancing optimality and local constraints. | • Less responsive to rapidly changing environments if prediction horizon is mismatched. • Highly sensitive to operating conditions and velocity range. • Challenges with coupled control dynamics; poor performance in tightly coupled maneuvers (e.g., Figure-Of-Eight), experiencing significant difficulties in completing the Figure-Of-Eight trajectory and exhibiting instability. | • General obstacle avoidance. • Path optimization in somewhat predictable dynamic environments. | • Highly dynamic environments with rapid changes. • Complex multi-axis maneuvers, scenarios where tuning is difficult. |
| Invariant-set | • Provides strong safety guarantees by design and ensures adherence to operational constraints. • Superior stability and performance in complex, tightly coupled maneuvers (e.g., Figure-Of-Eight). • Smooth control profile with low average acceleration and curvature. | • Intrinsically imposes a minimum turning radius, leading to fixed minimum invariant set size. • Can be inefficient or unsuitable in extremely constrained spaces (e.g., avoids very narrow passages). • May lead to non-optimal behavior from mission objectives in specific constrained tasks. | • Missions prioritizing safety and guaranteed constraint adherence. • Complex and continuous maneuvers. • Highly dynamic and cluttered environments. | • Tasks requiring navigation through extremely narrow passages. • Scenarios where strict path optimality (shortest path) is the sole objective. |
TABLE VII: Comparison of Controllers for UUV Navigation
More in this topic
Underwater Robotics
Underwater robotics develops autonomous and remotely operated machines for navigation, inspection, manipulation, and exploration in aquatic environments.
See all 147 papers