Artificial Landmarks for Precise Autonomous Vehicle Localization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing autonomous driving systems for passenger transport vehicles face challenges in achieving robust, precise, and cost-effective localization in mixed traffic environments, as conventional methods often require complex sensor systems and infrastructure modifications.
Innovation Solution
The method employs artificial landmarks with unique code markings, recognizable by sensors like laser scanners, which include calibration strips and optional QR codes, attached to existing infrastructure at a height above the vehicle to ensure simple, robust, and cost-effective localization, preventing damage and manipulation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional sensor-based localization methods are used, then localization capability is achieved, but device complexity and cost increase
Solution Approach 1:
The patent uses optical copies (visual appearance) of landmarks instead of complex sensor measurements. The mobile platform's sensor captures an image of the landmark, and the system determines position by recognizing the visual appearance and comparing it with stored reference images, replacing complex multi-sensor fusion with simple image recognition
Solution Approach 2:
The patent replaces complex mechanical/multi-sensor localization systems with a simple camera-based optical recognition system. Instead of using multiple independent sensors (lidar, radar, ultrasonic) and complex fusion algorithms, the system uses a single camera to capture optical information and determine position through image processing
2Measurement precision
If robust and precise localization is achieved, then positioning accuracy improves, but infrastructure cost increases
Solution Approach 1:
The patent uses inexpensive, simple landmarks that can be easily manufactured and deployed. These landmarks consist of basic visual markers (patterns, colors, shapes) that can be printed on standard materials, replacing expensive permanent infrastructure installations while achieving sufficient positioning accuracy
Solution Approach 2:
The patent designs landmarks with multiple visual features (geometric patterns, color codes, hierarchical structures) that serve multiple functions: identification, positioning, orientation, and calibration. This multi-functionality allows simple physical structures to provide comprehensive localization information without requiring multiple types of infrastructure
3Reliability
If landmarks are placed at elevated positions, then localization reliability improves, but vulnerability to damage decreases
Solution Approach 1:
The patent places landmarks in the vertical dimension (elevated on poles, buildings, or suspended) rather than only at ground level. This vertical displacement removes landmarks from the horizontal plane where vehicles and pedestrians operate, reducing physical interference and damage risks while maintaining optical visibility for localization
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enables precise and reliable localization of autonomous vehicles with reduced computing time and infrastructure costs, utilizing sensors with 360° views and avoiding interference from other road users, while ensuring sensor calibration and providing additional identifiability through varied bar configurations and QR codes.
Implementation Method 1
laser scanners on the mobile platforms... Laser scanners are already being used on mobile platforms in connection with, for example, distance measurements and also for localization
Data Source
Figure 1
Figure 2~3
Figure 4
AI summary
In a system of artificial landmarks (100) in an environment (20) for use for sensor-based localization of mobile platforms (10), the artificial landmarks (100) each have a code marking that serves to uniquely identify the individual landmarks (100).