Convolutional Neural Network Flight Path Deviation Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing flight route pattern recognition techniques are computationally inefficient and inaccurate, often resulting in high false positive or false negative rates, failing to effectively identify flight path deviations due to their reliance on static rule-based approaches that cannot accommodate the vast variations in flight paths.
Innovation Solution
The use of a convolutional neural network (CNN) algorithm to classify flight data by applying filters to subsets of coordinates corresponding to time intervals, enabling the identification of flight path deviations such as detours or holding patterns with high accuracy and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If rule-based approaches are used to compare actual distance travelled to planned distance at waypoints, then the system can identify flight path deviations, but the computational efficiency is poor and accuracy is low with false positive or false negative rates up to 55%
Solution Approach 1:
The patent replaces the mechanical rule-based comparison system with a neural network-based intelligent system. The CNN algorithm automatically learns and applies complex patterns for flight path deviation detection, substituting the manual rule-based mechanical approach with an adaptive intelligent system that achieves both high accuracy and computational efficiency.
Solution Approach 2:
The patent changes the fundamental parameters of the detection system by transitioning from static rule-based thresholds to dynamic neural network parameters. The CNN model learns optimal detection parameters from training data, enabling adaptive parameter adjustment that improves both accuracy and efficiency compared to fixed rule-based parameters.
2Adaptability or versatility
If rule-based approaches with multiple rules are implemented to accommodate vast variations in flight paths, then detection coverage improves, but computational complexity and processing time increase significantly
Solution Approach 1:
The patent implements a universal neural network model that can detect multiple types of flight path deviations through a single system. The CNN architecture is designed to handle various deviation patterns (detours, holding patterns, unexpected weather responses, traffic pattern adjustments) using the same core detection mechanism, eliminating the need for separate rule sets for each deviation type.
Solution Approach 2:
The neural network system performs self-learning and self-adjustment through training on historical flight data. The model automatically adapts to new flight path variations without requiring manual rule updates, enabling the system to serve itself in maintaining detection accuracy across diverse flight scenarios while keeping system complexity manageable.
3Reliability
If more rules are added to cover all flight path variations, then detection accuracy improves, but the system becomes computationally inefficient and harder to maintain
Solution Approach 1:
The patent performs preliminary training of the neural network model using historical flight data before deployment. This preliminary action allows the system to learn and internalize numerous flight path patterns in advance, so that during actual operation, the model can reliably detect deviations without requiring complex real-time rule evaluations, simplifying implementation and maintenance.
Data Source
AI summary
A method of providing an indication identifying a flight as including or not including a flight path deviation, the method includes receiving flight data associated with a flight, the flight data comprising a plurality of coordinates indicating a location of an aircraft at a plurality of times during the flight. The method also includes classifying a plurality of time intervals of the flight by applying a convolutional neural network algorithm to subsets of the plurality coordinates to determine that the time interval includes or does not include at least one flight path deviation. The method also includes identifying the flight as including at least one flight path deviation if at least one of the time intervals is classified as including the at least one flight path deviation and providing an indication identifying the flight as including at least one flight path deviation to a second computing system in response to identifying the flight as including at least one flight path deviation.