Neural Network Runway Occupancy Prediction System
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems for predicting the time of occupancy of a runway by an aircraft require high computing capacity and time, which is inefficient and may lead to uncertainties in air traffic management.
Innovation Solution
A method and system that predict the time of occupancy of a runway by generating current aircraft movement parameters, allowing crew members to select an exit type and distance, and using a neural network trained on landing data to compute the occupancy time, thereby reducing computational resources and time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If current systems use traditional computational methods to predict runway occupancy time, then prediction accuracy can be maintained, but computing capacity and time requirements become excessively high
Solution Approach 1:
The patent replaces traditional mechanical computational methods with a neural network-based system. The neural network is trained on historical landing data to learn complex patterns of runway occupancy behavior, substituting heavy mathematical computations with a pre-trained intelligent model that provides both high accuracy and fast prediction speeds during actual runway occupancy time estimation
Solution Approach 2:
The neural network is trained in advance using extensive historical landing data from multiple aircraft types and conditions. This preliminary training phase allows the system to pre-process and learn from large datasets, so that during actual runway operations, the prediction can be made quickly without requiring real-time heavy computation, thus resolving the contradiction between accuracy and computing efficiency
2Productivity
If ground-based air traffic control systems use statistical data for prediction, then computing resources are conserved, but prediction accuracy deteriorates due to lack of current aircraft parameters
Solution Approach 1:
The system enables the aircraft itself to perform the prediction computation through an on-board neural network processor. By transferring the computational task from ground-based air traffic control to the aircraft's own systems, the solution maintains high prediction accuracy using current aircraft parameters while the aircraft autonomously handles the computation, freeing ground systems for other air traffic management functions
3Reliability
If additional margins are added to separation distances to account for uncertainty, then safety is improved, but runway capacity and efficiency decrease
Solution Approach 1:
The system provides real-time feedback to air traffic control by continuously monitoring current aircraft parameters (speed, weight, configuration, weather conditions) and using the neural network to predict the actual runway occupancy time. This accurate, dynamic feedback replaces static conservative estimates, allowing air traffic control to optimize separation distances based on actual predicted performance rather than adding excessive safety margins, thus maintaining safety while improving runway capacity
Data Source
AI summary
A prediction system with a unit for generating a current value of a parameter relating to the movement of an aircraft, a selection unit for a crew member of the aircraft to select an exit of the runway used for landing, an avionics computer for computing a value of the time of occupancy of the runway and a unit for transmitting this value to a user system. The avionics computer using an initial ground speed and an exit type and an exit distance of the exit selected, and computes the value of the time of occupancy of the runway, with the data and by using a prediction model comprising a neural network trained on the basis of landing data originating from previous landings.