Vehicle Localization Using Visual Mile Marker Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Vehicle localization accuracy is compromised by errors in GPS and map data, particularly in environments where GPS signals are unreliable, such as bridges and tunnels, which can lead to incorrect lane determination and control issues in autonomous driving systems.
Innovation Solution
The use of vehicle cameras to detect features like mile markers, comparing visual data with map and GPS data to calculate localization errors, and applying these errors to correct the vehicle's position, thereby enhancing localization accuracy and providing redundancy to GPS-based determination.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If GPS and map data are used for vehicle localization, then the system is simple and provides global position information, but localization accuracy deteriorates due to biases and errors in the data
Solution Approach 1:
The patent combines GPS-based localization with visual feature detection (camera-based mile marker recognition) to create a hybrid localization system. The controller integrates data from both the navigation system and image sensor, comparing visual distance measurements with GPS-derived distances to correct localization errors and improve accuracy beyond what either system could achieve alone.
Solution Approach 2:
The system implements feedback by continuously comparing the distance to mile markers measured by the camera with the distance calculated from GPS coordinates and map data. When discrepancies are detected, the system uses this feedback to identify and correct localization errors, adjusting the vehicle's perceived position to align with actual physical features.
2Measurement precision
If visual feature detection is added to improve localization accuracy, then measurement precision improves, but device complexity increases due to additional sensors and processing
Solution Approach 1:
The controller serves multiple functions: it processes navigation data from the GPS system, analyzes image data from the camera to detect mile markers, calculates distances using both methods, compares results to identify errors, and applies corrections. This multi-functionality reduces the need for separate dedicated components for each task, managing complexity through consolidation.
Solution Approach 2:
The controller acts as an intermediary that reconciles data from two different localization approaches (GPS and visual detection). It mediates between the global positioning system and the camera-based measurement system, integrating their outputs to produce a corrected, more accurate localization result while managing the complexity of coordinating multiple data sources.
3Ease of operation
If GPS data is used for lane determination, then the process is simple, but reliability deteriorates in environments with unreliable GPS signals such as bridges and tunnels
Solution Approach 1:
The system performs preliminary visual detection of mile markers and road features to establish baseline position information before relying solely on GPS. By pre-acquiring visual reference data and comparing it with upcoming GPS readings, the system prepares correction factors in advance, ensuring reliable lane determination even when GPS signals become unreliable in challenging environments.
Data Source
AI summary
Methods and systems for vehicle localization are disclosed. An exemplary system includes a navigation system configured to generate navigation data corresponding to a global position of the vehicle, at least one image sensor configured to capture image data of a selected roadway feature along a projected path of the vehicle, a database comprising map data corresponding to lateral and longitudinal coordinates for a plurality of roadway features along the projected path of the vehicle; and a controller, the controller configured to receive the image data, the map data, and the navigation data, calculate a first distance from the selected feature to the vehicle using the navigation data and the image data, calculate a second distance from the selected feature to the vehicle using the navigation data and the map data, and determine a localization error by comparing the first distance to the second distance.