Neural Network Positioning System for Mobile Units
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing positioning systems for mobile units, such as vehicles, are susceptible to malfunctions and inaccuracies due to environmental conditions and component failures, which can hinder precise localization and autonomous control.
Innovation Solution
A positioning system that includes a reference localization unit, a capturing unit, and a computing unit utilizing a neural network to determine the mobile unit's position by comparing captured movement data with reference positions, allowing for continuous learning and adjustment of neural network parameters to improve redundancy and accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional positioning systems are used, then the system is simple to operate, but the reliability of position determination deteriorates due to environmental conditions and component failures
Solution Approach 1:
The positioning system is segmented into multiple independent components: a reference localization unit for capturing reference positions, a capturing unit for obtaining current position data, and a computing unit with a neural network for calculating position. This segmentation allows each component to be optimized independently and provides redundancy, improving reliability without requiring a completely complex integrated system.
Solution Approach 2:
A neural network serves as an intermediary between the reference localization unit and the capturing unit. The neural network processes and interprets the captured position data, enabling the system to determine position even when direct signal reception is restricted. This intermediary layer adds robustness against environmental conditions and component failures.
2Measurement precision
If neural network-based positioning is implemented, then the measurement precision of position determination is improved, but the device complexity increases due to the need for continuous learning and parameter adjustment
Solution Approach 1:
The neural network is pre-trained with learning data before actual positioning operations begin. This preliminary training phase allows the system to develop accurate position determination capabilities in advance, so that during operational use, the system can achieve high measurement precision without requiring continuous complex learning processes, thereby reducing operational complexity.
Solution Approach 2:
The neural network performs self-adjustment of its parameters based on the comparison between reference positions and calculated positions. This self-service capability allows the system to maintain and improve its own precision automatically without requiring external intervention or complex manual calibration procedures, balancing precision improvement with system simplicity.
3Reliability
If redundant positioning methods are used, then the reliability of autonomous control is improved, but the loss of processing time increases due to multiple calculation paths
Solution Approach 1:
The system implements a feedback mechanism where the reference localization unit continuously provides reference positions that are compared with positions calculated by the neural network. This feedback loop allows the system to verify and correct position determinations in real-time, improving reliability without requiring multiple independent calculation paths that would increase processing time.
Solution Approach 2:
The system uses partial redundancy by having the neural network operate primarily with captured position data, while the reference localization unit serves as a verification and correction mechanism. This partial redundancy approach provides sufficient reliability for autonomous control without the full overhead of multiple complete calculation paths, thereby minimizing processing time loss.
Data Source
AI summary
The invention relates to a positioning system for a mobile unit with a reference localization unit, through which a first reference position can be captured at a first point in time and a second reference position of the mobile unit can be captured at a later second point in time; a capturing unit, through with movement data of the mobile unit can be captured; and a computing unit, through which a calculated second position of the mobile unit can be determined at the second point in time by means of a neural network based on the first reference position and the captured movement data. At least one parameter of the neural network can thereby be adjusted based on a comparison of the second reference position with the calculated second position. The invention further relates to a method for operating a positioning system for a mobile unit, in which a first reference position of the mobile unit is captured at a first point in time. Furthermore, movement data from the mobile unit is captured, and a calculated second position of the mobile unit is determined at the second point in time by means of a neural network based on the first reference position and the captured movement data. A second reference position of the mobile unit is captured at the second point in time, wherein a parameter of the neural network is adjusted based on a comparison of the second reference position with the calculated second position.