Vehicle Visual Localization Using Reference Image Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous driving systems face challenges in accurately detecting and avoiding obstacles, particularly those not mapped on conventional roadmaps, such as potholes, due to limitations in current sensor technologies and processing methods.
Innovation Solution
A method involving a visual sensor that generates a signature of the vehicle's environment by comparing acquired images to reference images, using a multidimensional representation and iterative dimension expansion and merge operations to identify relevant features, thereby determining the vehicle's location with high resolution and power efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional sensor technologies and processing methods are used for obstacle detection, then the system structure remains simple, but the detection precision and ability to identify unmapped obstacles deteriorates
Solution Approach 1:
The patent transforms the 2D image data into a 3D spatial representation by introducing depth information through stereo vision or depth sensors. This dimensional transformation enables precise localization of unmapped obstacles by creating a three-dimensional map that incorporates height, width, and depth, thereby improving detection precision without proportionally increasing system complexity
Solution Approach 2:
The patent divides the visual field into multiple regions of interest and processes them separately through iterative dimension expansion. By segmenting the environment into discrete spatial cells and analyzing each independently, the system achieves high detection precision for individual obstacles while managing computational complexity through distributed processing
2Measurement precision
If high-resolution localization is achieved through detailed image processing, then the location accuracy improves, but the power consumption increases
Solution Approach 1:
The patent applies partial action by processing only the necessary portions of the visual data at high resolution. Through iterative dimension expansion, the system focuses computational resources on expanding specific spatial dimensions only when obstacles are detected or in critical areas, rather than processing the entire field of view at maximum detail, thereby achieving high location accuracy while controlling power consumption
Solution Approach 2:
The patent implements periodic action through iterative processing cycles where the dimension expansion is performed in repeated stages. The system alternates between scanning the environment, detecting changes, and performing detailed processing only when necessary, creating a periodic pattern of high-power and low-power operation that maintains accuracy while managing energy usage
3Reliability
If comprehensive environmental mapping is performed to detect all obstacles, then the obstacle detection capability improves, but the processing time increases
Solution Approach 1:
The patent applies preliminary action by pre-processing the visual data through dimension expansion before full obstacle detection. The system performs initial spatial transformation and identifies potential obstacle regions in advance, allowing the main detection algorithm to focus only on suspicious areas rather than analyzing every pixel, thereby improving detection reliability while reducing overall processing time
Solution Approach 2:
The patent implements skipping by rapidly processing certain dimensions or regions that are determined to be low-risk or already-mapped. Through iterative dimension expansion, the system identifies and skips over stable, well-understood areas of the environment, rushing through routine processing while dedicating more time to novel or hazardous regions, thus maintaining high detection reliability without excessive processing time
Data Source
AI summary
A method for determining a location of a vehicle, the method may include receiving reference visual information that represents multiple reference images acquired at predefined locations; acquiring, by a visual sensor of the vehicle, an acquired image of an environment of the vehicle; generating, based on the acquired image, acquired visual information related to the acquired image, wherein the acquired visual information comprises acquired static visual information that is related to the environment of the vehicle; searching for a selected reference image out of the multiple reference images, the selected reference image comprises selected reference static visual information that best matches the acquired static visual information; and determining an actual location of the vehicle based on a predefined location of the selected reference image and to a relationship between the acquired static visual information and to the selected reference static visual information; and wherein the determining of the actual location of the vehicle is of a resolution that is smaller than a distance between the selected reference image and a reference image that is immediately followed by the selected reference image.