Location Accuracy via Image-Based Voxel Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current GPS and cellular triangulation methods for ride-hailing services often result in positional inaccuracies, leading to delays and missed pickups, especially in urban environments with limited sight lines and environmental obstructions, which increases costs and reduces service efficiency.
Innovation Solution
A system that uses a processor to capture and process image data from a user's environment, generating a 3D point cloud using structure-from-motion algorithms, removing temporal objects through semantic segmentation, and converting it into a voxel map to accurately determine the user's location relative to known objects, and transmitting this data to improve location accuracy and suggest alternative pickup points.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If GPS and cellular triangulation methods are used to identify user location, then the system can obtain location information with minimal additional infrastructure, but the positional accuracy is limited and insufficient for reliable ride-hailing pickups
Solution Approach 1:
The patent introduces image data and environmental features as an intermediary to bridge the gap between GPS coordinates and actual physical location. By capturing images of the user's surroundings and identifying distinctive features (buildings, street signs, landmarks), the system creates a visual reference system that mediates between the imprecise GPS location and the precise pickup point, resolving the accuracy issue without requiring complex additional hardware infrastructure
Solution Approach 2:
The patent replaces the purely signal-based mechanical system (GPS/cellular triangulation) with a visual recognition system. Instead of relying solely on satellite signals and cellular tower triangulation, the system substitutes image processing and feature recognition to determine location, achieving higher precision by analyzing visual environmental cues rather than depending on the limited precision of traditional positioning methods
2Productivity
If the driver and user must visually identify each other through some method, then pickup can occur, but limited sight lines due to obstructions or environmental variables increase the time spent looking for the vehicle
Solution Approach 1:
The patent applies preliminary action by determining the precise pickup location in advance using image data analysis before the driver arrives. The system identifies the user's exact position relative to environmental features and communicates this information to the driver beforehand, eliminating the need for time-consuming visual searching when the driver arrives at the location
Solution Approach 2:
The system implements feedback by providing the driver with directional guidance and location confirmation based on analyzed environmental features. The driver receives information about what to look for (specific buildings, signs, or landmarks) and the user's precise location relative to these features, creating a feedback loop that guides the driver efficiently to the pickup point without random searching
3Reliability
If GPS location data is used without considering environmental context, then the system can operate with simple processing, but the user may be located on the wrong side of a corner or in a location that does not allow parking
Solution Approach 1:
The patent changes the parameters used for location determination from purely coordinate-based (latitude/longitude) to include visual environmental parameters. By analyzing image data for distinctive features, building orientations, street signs, and landmarks, the system transforms the location representation to include contextual information about the physical environment, enabling reliable determination of whether the user is on the correct side of the street or in a parking-permitted location
Solution Approach 2:
The patent adds a visual dimension to location determination by incorporating image data analysis alongside traditional GPS coordinates. This creates a multi-dimensional location system that combines numerical coordinates with visual environmental context, allowing the system to distinguish between locations that have similar GPS coordinates but different physical characteristics (such as opposite sides of a corner or parking availability)
Data Source
AI summary
A system includes a processor configured to request capture of image data of an environment surrounding the user, responsive to a margin of error of a detected location of a user being above a predefined threshold. The processor is also configured to process the image data to determine an actual user location relative to a plurality of objects, having known positions, identifiable in the image data and replace the detected location with the determined actual user location.