Neural Network Air Traffic Training Platform
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current air traffic control training systems lack the ability to effectively utilize experienced knowledge, offering limited scenarios and being non-upgradeable, making them inadequate for representing real-world air traffic control situations.
Innovation Solution
A training and assistance platform utilizing an electronic air traffic management system with a neural network derived from historical data to provide dynamic and realistic training scenarios, allowing air traffic controllers to receive instructions based on real-time data and compare their decisions with automatically determined instructions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional test programs are used for air traffic controller training, then training can be conducted with existing systems, but the number of scenarios is very limited and the system is not upgradeable
Solution Approach 1:
The training system transitions from static test programs to dynamic scenarios generated by a neural network that adapts to new air traffic control situations. The system continuously learns from historical data and operational experiences, automatically updating training scenarios without requiring manual system upgrades or reprogramming.
Solution Approach 2:
The neural network automatically generates and updates training scenarios by learning from historical air traffic control data and operational experiences. The system self-upgrades its training capabilities without external intervention, continuously improving scenario variety and realism based on accumulated knowledge.
2Reliability
If traditional test programs are used for training, then implementation is simple, but the programs are not representative of changes to air traffic control systems
Solution Approach 1:
The patent replaces traditional mechanical test program structures with a neural network-based system that processes air traffic control data. This substitution enables the system to automatically generate realistic training scenarios that reflect actual air traffic control situations and system changes, significantly improving representativeness.
Solution Approach 2:
The neural network creates accurate copies of real air traffic control situations by learning from historical data and operational experiences. These copied scenarios faithfully represent actual air traffic control environments, providing realistic training without requiring direct access to live systems.
3Adaptability or versatility
If a neural network is implemented for automatic instruction determination, then training scenarios become dynamic and representative of real situations, but system complexity increases
Solution Approach 1:
The neural network serves multiple functions simultaneously: it determines training scenarios, generates instructions, evaluates controller performance, and continuously learns from new data. This multi-functionality reduces the need for separate systems while maintaining high training scenario realism and adaptability.
Data Source
AI summary
A training and/or assistance platform for air traffic management is provided. The platform includes an air traffic management electronic system for obtaining input data representative of air traffic, to deliver, to an air traffic controller, information established as a function of the obtained input data, and to receive instructions from the air traffic controller The platform further includes a block for automatically determining instructions based on input data representative of at least the state of air traffic. The platform further includes an electronic processing module for collecting said input data and to provide it to the automatic determining block The platform further includes a neural network derived from learning on an input data history obtained by an electronic air traffic control system and received air traffic control instruction(s) received by the system.