Virtual Rail Vehicle Guidance Using Ground Signature Localization

Resolve Bottlenecks,
Find Innovative Solutions
Generate 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

VSEngineering 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

Engineering Contradiction:
Improveself-localization reliabilityVSAvoidnavigation system complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

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.

Inventive Principle:
Principle #2Taking out (Extraction)

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.

Inventive Principle:
Principle #26Copying

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

Engineering Contradiction:
Improvefeature detection adaptabilityVSAvoidfeature detection reliability
Core Design Contradiction:
Adaptability or versatilityVSReliability

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.

Inventive Principle:
Principle #13The other way round (Inversion)

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.

Inventive Principle:
Principle #33Homogeneity

3Measurement precision

If signature matching is performed for position determination, then the vehicle can be localized, but computational effort increases with environmental variations

Engineering Contradiction:
Improveposition determination precisionVSAvoidcomputational processing power
Core Design Contradiction:
Measurement precisionVSPower

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.

Inventive Principle:
Principle #1Segmentation

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.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP3482622B1Method for automatically guiding a vehicle along a virtual rail system
Publication Date: 2024.06.26 ROBERT BOSCH GMBH
  • EP3482622B1 patent drawingFigure 1~2
  • EP3482622B1 patent drawingFigure 3
  • EP3482622B1 patent drawingFigure 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.