Rail Vehicle Visual Odometry Localization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for localizing rail vehicles using GPS or global navigation satellite systems are inaccurate and unreliable due to potential errors or unavailability of signals, necessitating a more precise and independent localization method.
Innovation Solution
A method utilizing visual odometry based on image data from a camera sensor, which determines the position of the rail vehicle by identifying characteristic elements and accounting for movement models, allowing for precise localization independent of satellite data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If GPS or global navigation satellite systems are used for localization, then the localization can be performed with conventional navigation systems, but the localization accuracy and reliability deteriorate due to signal faults or unavailability
Solution Approach 1:
The patent introduces visual odometry as an intermediary method between the rail vehicle and the final localization result. Instead of directly relying on satellite signals, the system uses camera sensors to capture images of characteristic elements (such as track features, landmarks, or artificial markers) and processes these images through visual odometry algorithms to compute the vehicle's position and orientation. This intermediary approach provides reliable localization even when satellite signals are unavailable or faulty.
2Measurement precision
If visual odometry is performed using multiple characteristic elements in multiple frames, then the localization precision improves, but the computational requirements and processing time increase
Solution Approach 1:
The patent performs preliminary actions by pre-identifying and tracking characteristic elements across multiple image frames before final position calculation. The system continuously captures images, detects characteristic elements in each frame, and maintains their positions over time. This preliminary processing of image data and element tracking prepares the necessary information in advance, enabling accurate localization computation without excessive processing delays when results are needed.
Solution Approach 2:
The system dynamically adjusts its processing by continuously updating the positions of characteristic elements as new image frames arrive. Instead of static processing, the visual odometry method adapts to changing scenes by tracking moving elements and adjusting calculations based on the most recent image data. This dynamic approach maintains precision while optimizing processing efficiency through continuous rather than batch processing.
Data Source
Figure 1~4
AI summary
The invention relates to a method (100) for locating a rail vehicle (301), comprising: receiving (101) image data from a camera sensor (303) of the rail vehicle (301), wherein the image data depicts the environment of the rail vehicle (301) and wherein the image data comprises a plurality of frames recorded at different times; performing (103) a visual odometry procedure based on the image data; and determining (105) a position of the rail vehicle (301) based on the results of the visual odometry procedure. The invention further relates to a method for controlling a rail vehicle (301).