Parking Lot Detection from Truncated Probe Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods fail to accurately analyze probe data to determine the location, size, classification, or other details of parking lots in digital maps due to data truncation for privacy reasons, which is crucial for navigation systems, especially in urban areas where parking information is frequently outdated.
Innovation Solution
The method involves identifying patterns in probe data to detect parking lot locations, clustering relevant data points, using Principal Component Analysis to determine the shape and size of parking lots, and integrating them into the digital map, along with classifying and attributing them, while connecting them to the road network.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If probe data is truncated for privacy reasons, then user privacy is protected, but the ability to accurately detect parking lot locations and characteristics is degraded
Solution Approach 1:
The patent introduces an intermediary processing layer that analyzes probe data patterns without requiring access to complete raw trajectories. By using aggregated statistical features and pattern recognition on truncated data, the system mediates between privacy protection and detection accuracy, extracting meaningful parking lot information from limited data while preserving user privacy.
Solution Approach 2:
The patent creates simplified representations (copies) of parking lot characteristics from truncated probe data. Instead of requiring complete trajectory information, the system generates approximate location, size, and classification data that captures essential parking lot attributes without needing the full original data, enabling accurate detection from privacy-truncated inputs.
2Loss of information
If traditional methods are used to update digital maps, then map data remains static and outdated, but the complexity and cost of manual updates increase
Solution Approach 1:
The patent enables the digital map system to automatically update itself by processing probe data from user devices. The system self-updates parking lot locations, sizes, and classifications by analyzing patterns in probe data, eliminating the need for manual surveying and map editing. This self-service approach keeps map data current while reducing operational complexity and costs.
Solution Approach 2:
The patent implements a feedback loop where probe data from navigation devices continuously flows back to the map system, which then updates map information and makes it available for subsequent routing decisions. This closed-loop feedback mechanism ensures map data remains synchronized with real-world conditions without requiring active manual intervention.
3Measurement precision
If probe data from multiple sources is collected to improve parking lot detection accuracy, then detection precision improves, but data processing complexity and computational requirements increase
Solution Approach 1:
The patent merges probe data from multiple sources and devices into unified parking lot detections. By combining trajectories and stop patterns from numerous independent probes, the system achieves high detection accuracy through aggregated evidence. The merging process consolidates redundant information and uses statistical methods to resolve conflicts, managing complexity through systematic data integration rather than processing each probe independently.
Solution Approach 2:
The patent segments the complex task of parking lot detection into distinct analytical components: trajectory segmentation to identify stop sequences, spatial segmentation to define parking lot boundaries, and temporal segmentation to filter relevant probe data. This segmentation breaks down the complex processing task into manageable stages, reducing overall computational complexity while maintaining detection accuracy.
Data Source
Figure 1
Figure 2~3
Figure 4~5
AI summary
Parking lot (18, 20) locations and geometries are discerned from probe data (22) and integrated with a digital map (12) for use in navigation and other map-related activities. The steps of this invention include: identifying the positions in the probe data (22) where cars are probably parked, detecting the positions of parking lots (18, 20), determining the extension and shape of the detected parking lots (18, 20), and adjusting the parking lots (18, 20) into the road network (14). Optionally, the parking lot (18, 20) can be classified and additional attributes computed. Finally, a topological connection of the parking lot (18, 20) is made to the road network (14) during which entrances (34) and exits (36) are identified.