Vehicle-Passenger Recognition for Accurate Crowded Pickups
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Passengers and drivers face difficulties in identifying each other in crowded environments, especially with autonomous vehicles, as traditional methods rely on visual recognition which can be unreliable and distracting, and GPS coordinates are not precise enough.
Innovation Solution
A vehicle system with a camera and processor for identifying passengers using coarse and fine passenger attribute information, and a passenger device for identifying vehicles using vehicle identification information, enabling bi-directional recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional visual recognition methods are used for passenger identification, then the process is simple and does not require additional equipment, but the reliability is low and causes driver distraction in crowded environments
Solution Approach 1:
The patent replaces the mechanical/visual recognition system (driver visually identifying passenger) with an optical sensing system (camera capturing images) combined with digital image processing and pattern recognition algorithms. The camera captures passenger images, which are then processed through algorithms that compare extracted features against stored passenger profiles, automatically identifying the passenger without requiring driver visual attention.
Solution Approach 2:
The patent introduces an intermediary identification system consisting of cameras, image processing units, and pattern recognition software that mediates between the passenger and the vehicle. This intermediary system handles the complex task of passenger identification, allowing the driver to remain focused on the road while the system automatically verifies passenger identity through multiple attributes such as facial features, clothing, and physical characteristics.
2Measurement precision
If GPS coordinates are used for location identification, then the system is simple to implement, but the precision is insufficient in crowded environments
Solution Approach 1:
The patent segments the location identification process into multiple hierarchical levels: first using GPS coordinates for broad area localization, then progressively refining precision through street-level imagery matching, building recognition, and finally specific landmark or door-level identification. This segmentation allows the system to achieve high precision by combining multiple detection methods at different scales rather than relying on a single GPS coordinate.
Solution Approach 2:
The patent transitions from two-dimensional GPS coordinate identification to multi-dimensional identification by incorporating visual dimensions (camera images of passenger and environment), temporal dimensions (sequence of images and movements), and attribute dimensions (clothing, physical characteristics). This dimensional expansion enables precise location and identity verification in crowded environments where GPS alone is insufficient.
3Measurement precision
If camera scanning and image processing are implemented for passenger identification, then identification accuracy improves in crowded environments, but the processing time increases
Solution Approach 1:
The patent performs preliminary actions by pre-processing and storing multiple attributes of passengers (facial features, clothing characteristics, physical build) in a database before the actual identification event. When a passenger needs to be identified, the system retrieves pre-stored profiles and compares them against current camera images, significantly reducing processing time compared to analyzing all possible attributes in real-time from scratch.
Solution Approach 2:
The patent employs partial action by selectively analyzing only the most distinctive and readily detectable passenger attributes first (such as overall appearance, clothing color, and prominent features) to quickly narrow down potential matches. Only if multiple candidates remain does the system proceed to more detailed and time-consuming analysis of finer features, thus reducing average processing time while maintaining high accuracy.
Data Source
Figure 1
Figure 2
Figure 3A
AI summary
A vehicle is configured to receive, from a passenger, a pickup request including an approximate location of the passenger, to scan for the passenger after arriving at the approximate location of the passenger, and to determine whether the passenger has been identified by comparing passenger attribute information to results of the scan, and to transmit an approximate location of the vehicle and vehicle identification information to the passenger when the passenger has not been identified or is not accessible for pickup. The passenger is picked up by the vehicle when the passenger has been identified and is accessible for pickup. In addition, the passenger may also use a portable electronic device to identify the vehicle based on the received approximate location of the vehicle and vehicle identification information.