NOTAM Message Correction Using LSTM Parsing and Field Validation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing NOTAM-messages often contain errors or inaccuracies due to varying quality and accuracy across different aviation authorities, leading to potential safety risks and operational disruptions in flight operations.
Innovation Solution
A computer-implemented method using a trained machine-learning algorithm to correct NOTAM-messages by identifying and selecting preferred values for traffic, purpose, scope, and NOTAM-code-components, utilizing a single multi-output long short-term memory neural network for parsing and pre-processing to enhance accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If NOTAM-messages are manually encoded by hand, then the flexibility and adaptability of the system is maintained, but errors and incomplete information occur leading to reduced reliability
Solution Approach 1:
The patent replaces manual mechanical encoding processes with an automated machine learning system. A trained machine learning algorithm automatically corrects NOTAM-messages by identifying and correcting errors in fields such as Q-field and E-field, substituting the manual encoding process with an automated intelligent system that improves accuracy without requiring human intervention for each message.
Solution Approach 2:
The system enables self-service correction of NOTAM-messages through the machine learning algorithm that automatically identifies and corrects errors without human assistance. The algorithm learns from training data and autonomously corrects messages, making the system self-sufficient for the correction task while maintaining high reliability.
2Reliability
If a machine learning algorithm is used to correct NOTAM-messages, then accuracy and reliability are improved, but the device complexity increases
Solution Approach 1:
The patent applies preliminary action by training the machine learning algorithm in advance using a comprehensive training set of NOTAM-messages before deployment. The algorithm learns from pre-prepared training data containing various error types and correct patterns, enabling it to perform accurate corrections without requiring complex real-time processing logic during operation.
3Object-affected harmful factors
If manual encoding is used for NOTAM-messages, then system simplicity is maintained, but safety risks and operational disruptions increase
Solution Approach 1:
The patent replaces manual encoding with automated machine learning-based correction to eliminate safety risks associated with human error. The system automatically processes and corrects NOTAM-messages, ensuring accurate information is provided to pilots and air traffic controllers without the safety hazards of manual encoding errors.
Data Source
AI summary
A computer-implemented method for correcting initial NOTAM-messages using a trained machine-learning algorithm, comprising: inputting a plurality of initial NOTAM-messages from a plurality of data sources, identifying, from the E-field and using the trained machine-learning algorithm, a plurality of permissible values for the traffic, the purpose, scope, first NOTAM-code-component, and second NOTAM-code-component; selecting a preferred value for one or more of the traffic, purpose, scope, first NOTAM-code-component, and second NOTAM-code-component, wherein selecting of the preferred value is done using a set of probabilities; and setting the selected value for the traffic, purpose, scope, first NOTAM-code-component, and second NOTAM-code-component as corrected values for the traffic, purpose, scope, first NOTAM-code-component, and second NOTAM-code-component to produce corrected NOTAM-message. The initial NOTAM-messages comprises at least a Q-field and an E-field. The E-field comprises a string describing a content of the initial NOTAM-messages.