Virtual Rail Vehicle Guidance Using Ground Signature Localization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing autonomous vehicle navigation systems face challenges in reliably guiding vehicles along predetermined paths, especially in environments with unstable or hard-to-detect features, and require efficient self-localization methods that can handle variations in surface conditions.
Innovation Solution
The method involves recording signal measuring points from a vehicle's sensor signals, forming sub-templates, and creating signatures from these measurements, which are then used to establish a virtual rail system with reference signatures stored in a correspondence table, allowing the vehicle to determine its position by matching working signatures with reference signatures for precise navigation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional SLAM methods are used for self-localization, then the vehicle can build maps and localize itself, but the system becomes complex and computationally intensive
Solution Approach 1:
The patent extracts the essential localization function from complex SLAM systems by using only ground surface pattern recognition. Instead of implementing full Simultaneous Localization and Mapping, the system extracts and processes only the necessary ground texture information to determine vehicle position, significantly reducing system complexity while maintaining localization reliability.
Solution Approach 2:
The patent creates simplified copies of ground surface patterns as templates for comparison. Rather than maintaining complex spatial maps, the system captures and stores representative ground surface images as template patterns, which are then matched against current sensor data to determine position, reducing computational requirements while preserving localization accuracy.
2Adaptability or versatility
If far-field structures are used for localization, then features can be detected from different angles, but these features are unstable and cannot be perceived under certain lighting conditions
Solution Approach 1:
Instead of using far-field structures as traditionally done, the patent inverts the approach by using near-field ground surface patterns immediately beneath the vehicle. This ground-level perspective provides stable, consistent features that are always visible regardless of lighting conditions or vehicle orientation, reversing the traditional localization paradigm while improving reliability.
Solution Approach 2:
The patent uses ground surface patterns that exhibit consistent homogeneous characteristics across different viewing conditions. The ground surface provides uniform texture and pattern properties that remain detectable under various lighting conditions, unlike the heterogeneous far-field structures that may become invisible or unstable under certain conditions.
3Measurement precision
If signature matching is performed for position determination, then the vehicle can be localized, but computational effort increases with environmental variations
Solution Approach 1:
The patent segments the ground surface into discrete template regions and processes each segment independently. By dividing the continuous ground surface into manageable template sections, the system can perform efficient pattern matching on smaller data units, reducing overall computational effort while maintaining position determination precision through cumulative template matching results.
Solution Approach 2:
The patent performs partial signature matching by comparing only critical template regions rather than processing the entire ground surface. By focusing computational resources on the most discriminative template areas that provide sufficient localization information, the system achieves accurate position determination with reduced computational power requirements.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
The invention relates to a method for automatically guiding a vehicle along a virtual rail system, wherein features of a surface over which the vehicle moves or will move are detected and converted into at least one working signature (A*-O*), wherein it is checked whether the at least one working signature (A*-O*) matches at least one reference signature (AO) of the virtual rail system, wherein the at least one reference signature (AO) is assigned a position on the virtual rail system, and, if the at least one working signature (A*-O*) and the at least one reference signature (AO) match, the position of the vehicle on the virtual rail system is inferred.