Vision Sensor Localization Using Encoded Road Sign Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current autonomous driving localization methods using Lidar/Camera with high-definition 3D map data are costly and face uncertainties due to environmental changes, necessitating a more cost-effective and reliable approach.
Innovation Solution
Implementing a vision sensor system that reads encoded data from road signs, providing structured data for localization, which includes a sensor block to detect and process data from infrastructure elements like road signs, and a processor to calculate distances and provide localization information for autonomous vehicles.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If Lidar/Camera with high definition 3D map data is used for localization, then positioning accuracy is improved, but cost increases
Solution Approach 1:
The patent replaces expensive Lidar systems with standard vision sensors (cameras) that are significantly cheaper and more widely available. The system uses encoded data sets displayed on road infrastructure elements as a temporary, disposable information source rather than relying on costly permanent sensing infrastructure.
Solution Approach 2:
The patent substitutes the mechanical/optical Lidar system with a vision-based computational system. Instead of using active light emission and time-of-flight measurements, the system uses passive image capture and computational processing of encoded visual patterns to achieve localization.
2Measurement precision
If Lidar/Camera with high definition 3D map data is used for localization, then positioning accuracy is improved, but reliability deteriorates due to environmental uncertainties
Solution Approach 1:
The patent changes the fundamental parameters of the localization approach by transitioning from geometric and intensity-based Lidar/Camera fusion to a vision-only system that relies on encoded information content. The encoded data sets contain explicit localization information that is invariant to environmental conditions, making the system more reliable across varying weather, lighting, and seasonal conditions.
3Ease of manufacture
If vision sensor is used to read road signs, then cost is reduced, but measurement precision may deteriorate
Solution Approach 1:
The patent applies preliminary action by pre-encoding localization data into the visual patterns on road infrastructure elements before the vehicle arrives. The encoded data sets contain pre-calculated positioning information that the vision sensor can directly decode, eliminating the need for complex real-time 3D map matching and uncertainty processing.
Solution Approach 2:
The patent introduces an intermediary - the encoded data set displayed on road infrastructure elements - that mediates between the vision sensor and the localization algorithm. This intermediary carries structured localization information that simplifies the sensing-to-positioning pipeline, enabling accurate localization with standard vision hardware.
Data Source
AI summary
A system having an encoded data set and a sensor. The encoded data set may be configured to store a plurality of information relating to a surrounding area. The encoded data set is presented in a vision sensor readable format along with human readable information on an infrastructure element. The sensor may be configured to (i) locate the encoded data set and (ii) calculate a distance to the road sign based on the information relating to the surrounding area. The distance is used to provide localization in an autonomous vehicle application.