Vessel Course Adherence Prediction Using Machine Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing navigation monitoring systems face challenges in accurately predicting whether a vessel navigates along a course, particularly in curved course scenarios, leading to potential collisions and erroneous risk determinations.
Innovation Solution
A navigation monitoring program utilizing a machine learning-based prediction model that integrates direction and position information of vessels to predict course adherence, incorporating features like course over ground, heading, and block IDs to enhance prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional navigation monitoring methods are used to determine course adherence, then the system is simple to operate, but the prediction accuracy deteriorates in curved course scenarios
Solution Approach 1:
The patent replaces traditional rule-based navigation monitoring methods with a machine learning model that uses neural networks to predict course adherence. The model processes direction information and position information to generate predictions about whether a vessel will navigate along its course, substituting mechanical decision rules with learned patterns from historical data.
Solution Approach 2:
The patent transforms navigation monitoring from a static rule-based system to a dynamic predictive system by changing the parameters processed (direction information, position information, block IDs) and the method of processing (machine learning inference). This allows the system to adapt to curved course scenarios by learning from historical navigation patterns.
2Measurement precision
If machine learning-based prediction model is implemented, then prediction accuracy improves, but computational resources and processing time increase
Solution Approach 1:
The patent performs preliminary action by pre-training the machine learning model offline using historical navigation data. The model learns patterns of course adherence during curved navigation beforehand, so that during actual navigation monitoring, predictions can be made quickly using the pre-learned knowledge stored in the model parameters.
Solution Approach 2:
The patent uses copying by training the model on historical data that copies real navigation scenarios. The model learns from replicated patterns in the training data, allowing it to generalize to new situations without requiring extensive real-time computation.
3Reliability
If more navigation data parameters are collected and processed, then prediction reliability improves, but data processing complexity increases
Solution Approach 1:
The patent applies universality by designing a single machine learning model that processes multiple types of navigation data (direction information, position information, block IDs) through a unified architecture. The model serves multiple functions: predicting course adherence, identifying curved course scenarios, and providing risk assessments, all through one versatile system.
Solution Approach 2:
The patent uses an intermediary approach by introducing a machine learning model as a mediator between raw navigation data and prediction results. The model transforms complex multi-parameter input data into simplified prediction outputs, acting as an intermediary layer that handles the complexity internally while presenting a clean interface for decision-making.
Data Source
AI summary
A non-transitory computer-readable storage medium storing a navigation monitoring program that causes at least one computer to execute a process, the process includes acquiring direction information that indicates a direction of a vessel and position information that indicates a position of the vessel; and predicting whether or not the vessel navigates along a course by inputting the acquired direction information and the acquired position information to a prediction model generated by machine learning by using direction information and position information for each of a plurality of vessels that has navigated in the past and a correct answer label that indicates whether or not each of the plurality of vessels navigates along a course.