Autonomous Vehicle Parking Verification via Visual Reference Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous driving vehicles lack the ability to independently determine if they are parked at a correct reference parking location, as they do not have a vehicle occupant to observe the surroundings.
Innovation Solution
A vehicle parking finder support system that uses a combination of cameras and object recognition processing to compare image data of the vehicle's surroundings with stored reference image data, determining if the vehicle is at the correct parking location based on identified objects and their attributes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If a vehicle occupant manually observes surroundings to verify parking location, then parking accuracy can be maintained, but the system cannot be applied to autonomous vehicles without occupants
Solution Approach 1:
The system enables the autonomous vehicle to perform its own parking verification independently without human intervention. The vehicle captures images of surroundings, automatically processes them through object recognition, and compares identified objects with pre-stored reference data to determine if it is parked at the correct location.
Solution Approach 2:
The manual observation and verification process performed by the vehicle occupant is replaced with an automated computer vision system. Cameras capture visual data, object recognition algorithms process the images to identify relevant objects and features, and automated comparison logic determines parking accuracy without mechanical human involvement.
2Ease of operation
If GPS is used to determine vehicle location, then positioning capability is provided, but accuracy is insufficient near buildings, trees, or underground areas
Solution Approach 1:
The system combines multiple positioning approaches by integrating visual recognition of environmental features with GPS coordinates. Object recognition identifies distinctive objects and features in the captured images, and this visual location information is merged with GPS data to achieve accurate parking verification even in areas where GPS alone is insufficient.
Solution Approach 2:
Visual features and objects captured in the environment serve as intermediary reference points between the vehicle and the parking location verification system. Instead of relying solely on direct GPS signals, the system uses recognized objects (buildings, trees, signs, etc.) as intermediate references to determine if the vehicle is at the correct parking location.
3Reliability
If reference image data is stored for comparison, then parking location verification is enabled, but the system must handle changes in surroundings over time
Solution Approach 1:
The reference data system is made dynamic rather than static. The system stores multiple sets of reference image data captured at different times, and the object recognition process adapts to identify consistent objects across temporal changes. This allows the system to maintain reliable verification accuracy while adapting to environmental changes such as seasonal variations, construction, or moved objects.
Data Source
Figure 1a
Figure 1b
Figure 2a~2c
AI summary
The disclosure relates to a vehicle parking finder support system (100), method and computer program product (500) for determining if a vehicle (1) is at a reference parking location, the vehicle parking finder support system (100) comprises at least a first camera (30a,30b,30c,30d) configured to provide images of the surroundings of the vehicle (1); a memory (101a,101b,101c); a processing circuitry (102a,102b,102c) operatively connected to the at least first camera (30a,30b,30c,30d) and the memory (101a,101b,101c), configured to cause the vehicle parking finder support system (100) to: obtain, by the at least first camera (30a,30b,30c,30d), at least first image data (im1,im2,im3) of the surroundings of the vehicle (1); obtain, from the memory (101a,101b,101c), at least first reference image data (rim1,rim2,rim3); and determine if the vehicle (1) is at the reference parking location using the at least first reference image data (rim1,rim2,rim3) and the obtained at least first image data (im1,im2,im3).