Autonomous Submarine Navigation Using Onboard Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Submarines and submersibles face challenges in navigating underwater environments due to the inability to use surface navigation techniques and the difficulty in detecting and discriminating between trawlers and other underwater obstacles, which can lead to safety risks and operational hazards.
Innovation Solution
A machine learning algorithm is trained onboard the watercraft using sensor signals, such as SONAR, to recognize patterns and maneuvers of deterrents like trawlers, allowing for autonomous navigation and avoidance of obstacles without external communication, utilizing techniques like dead reckoning, inertial navigation, and hydroacoustic sensing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If surface navigation techniques (satellite navigation, RADAR navigation) are used by submarines, then navigation accuracy is improved, but the submarine can be detected by enemy sensors and security is compromised
Solution Approach 1:
The patent extracts the navigation function from surface-based detection-prone systems (satellite, RADAR) and implements it using underwater-compatible inertial navigation and dead reckoning techniques. The submarine uses onboard sensors (gyrocompass, speed logs, fathometer) to determine position without emitting detectable signals or relying on surface infrastructure.
Solution Approach 2:
The patent introduces passive acoustic sensing as an intermediary method for detecting surface vessels. Instead of using active SONAR that emits detectable pulses, the system uses passive hydrophones to listen for acoustic signatures of trawlers and other surface vessels, enabling detection without revealing the submarine's position.
2Measurement precision
If active SONAR is used for navigation, then position determination is improved, but the submarine becomes detectable by enemy sensors
Solution Approach 1:
The patent removes active SONAR from the navigation system and replaces it with passive acoustic sensing combined with inertial navigation. Position is determined through dead reckoning using gyrocompass course information, speed logs, and fathometer depth measurements, supplemented by passive listening for external acoustic cues.
Solution Approach 2:
Passive acoustic sensing serves as an intermediary that provides environmental awareness without active emission. The system uses passive hydrophones to detect acoustic signatures of surface vessels and navigate around them, achieving situational awareness without the detection risk of active SONAR.
3Difficulty of detecting and measuring
If traditional sensor systems are used to detect deterrents, then detection capability is limited, but discrimination between trawlers and other obstacles is not achieved
Solution Approach 1:
The patent applies dynamic maneuvering patterns to differentiate trawlers from other vessels. The system detects and responds to specific maneuver characteristics (zigzag patterns, speed changes, heading adjustments) that are typical of trawling operations, enabling discrimination based on behavioral dynamics rather than static sensor readings.
Solution Approach 2:
The system continuously monitors acoustic signatures and maneuver patterns over extended periods, accumulating data on vessel behavior. This continuous observation enables reliable discrimination between trawlers and other obstacles by identifying consistent operational patterns rather than relying on transient or isolated sensor measurements.
4Reliability
If autonomous control is implemented, then navigation safety is improved, but the system complexity increases
Solution Approach 1:
The patent implements a multi-functional control system where a single autonomous controller integrates navigation, obstacle detection, deterrent identification, and evasive maneuvering. The ML algorithm serves multiple purposes: classifying vessel types, predicting trajectories, determining deterrent status, and generating avoidance commands, reducing the need for separate specialized systems.
Solution Approach 2:
The autonomous system performs self-learning and self-adjustment through machine learning algorithms that continuously analyze sensor data and improve deterrent classification accuracy over time. The system autonomously adjusts navigation decisions based on learned patterns without requiring external intervention or complex manual control systems.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The trained machine learning algorithm enables effective autonomous navigation of submarines and submersibles away from deterrents, enhancing safety and operational efficiency by distinguishing between different types of underwater structures and marine life, even in communication-isolated conditions.
Implementation Method 1
detection of trawls is generally not possible while discrimination between trawlers and other surface water craft may be problematic. Hence, the inventors have determined that autonomous control of navigation of submersed submarines and/or submersibles requires onboard machine learning. Particularly, the inventors have determined that deterrents such as trawlers may be characterised using sensor signals, for example relating to manoeuvres of the deterrents and/or using SONAR.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method of training a machine learning, ML algorithm to control a watercraft is described. The watercraft is a submarine or a submersible submerged in water. The method is implemented, at least in part, by a computer, comprising a processor and a memory, aboard the watercraft. The method comprises: obtaining training data including respective sets of sensor signals, related to respective deterrents, and corresponding actions of a set of communicatively isolated watercraft, including a first watercraft; and training the ML algorithm comprising determining relationships between the respective sets of sensor signals and the corresponding actions of the watercraft of the set thereof. A method of controlling a watercraft by a trained ML algorithm is also described.