Direction Finding Using Neural Networks to Resolve Ghost Locations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing direction finding methods face challenges in accurately locating transmitters due to 'ghost locations' that resemble actual transmitters, making automated and resource-efficient location determination difficult.
Innovation Solution
A method and system utilizing artificial neural networks, specifically pre-trained on graphical representations of bearings, to efficiently estimate transmitter locations, even in the presence of ghost locations, by separating the problem into bearing calculation and location estimation tasks, and using specialized neural networks for pattern recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If automated direction finding is performed using bearing data alone, then the process can be automated, but ghost locations occur that reduce measurement precision
Solution Approach 1:
The patent introduces an intermediary representation called a 'bearings picture' that transforms bearing data into a visual format. This intermediary allows automated analysis while eliminating ghost locations, as the visual representation makes it clear where actual transmitters are located versus where ghost locations appear. The bearings picture serves as a mediator between the automated system and the location estimation task.
Solution Approach 2:
The patent creates a visual copy or representation of the bearing data in the form of a bearings picture. Instead of directly processing raw bearing data which produces ghost locations, the system creates a graphical copy where bearings are displayed as lines or vectors on a map, allowing automated interpretation without the ghost location problem.
2Measurement precision
If graphical representation of bearings is output for operator estimation, then measurement precision can be maintained, but the process requires manual operation and loses automation
Solution Approach 1:
The patent enables the system to automatically interpret its own graphical output. The bearings picture is not just displayed for operator review but is automatically analyzed by a computer algorithm that can identify transmitter locations from the visual representation, allowing the system to serve itself in the location estimation task without human intervention.
Solution Approach 2:
The patent replaces the manual visual inspection process (mechanical human operation) with an automated computer-based image analysis system. The computer automatically processes the bearings picture to identify transmitter locations, substituting the mechanical human eye and brain with electronic image processing algorithms.
3Productivity
If specialized artificial neural networks are used for location estimation, then productivity increases, but device complexity increases
Solution Approach 1:
The patent applies preliminary action by pre-training the artificial neural network with bearing data and corresponding location information before actual use. The network is trained offline on a database of bearing pictures and known transmitter locations, so that during operation it can quickly and accurately estimate locations without requiring complex real-time processing. This pre-preparation reduces the complexity of the operational system.
Data Source
Figure 1a~1b
Figure 1c~1d
Figure 2a~2b
AI summary
A method for direction finding of at least one stationary and/or moving transmitter (T) comprises the following steps: a) measuring the signals emitted by each of the at least one transmitter (T) at at least two different measurement points (Pi, P2, P3), b) determining the location of the measurements points (Pi, P2, P3) at the time of the measurement, c) determining the bearings (26) from the measurement points (Pi, P2, P3) to each of the at least one transmitter (T), d) transferring the bearings (26) to a pre-trained artificial neural network (20), and e) estimating the locations (LT; LT1, LT2) of the at least one transmitter (T) by the artificial neural network (20). Further, a system (10) for direction finding of at least one stationary and/or moving transmitter (T) is shown.