Parking Lot Zone Estimation Using Grid-Based Parking Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining parking lot sections from vehicle data are costly and lack accuracy, with techniques like FCD analysis requiring manual map creation and open data sources like OSM having low coverage and maintenance issues.
Innovation Solution
A parking lot section estimation device and method that divides a region into grid cells, calculates feature amounts such as parking frequency and duration, and performs clustering to accurately identify parking lot sections without manual intervention.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual map creation is used to determine parking lot sections, then accuracy can be improved, but cost and complexity increase significantly
Solution Approach 1:
The system enables automatic self-determination of parking lot sections by having vehicles themselves provide the data needed for section identification through their GPS positional information and parking behavior patterns, eliminating the need for manual map creation while maintaining accuracy
Solution Approach 2:
The patent replaces manual mechanical map creation processes with automated computational methods that process vehicle GPS data through clustering algorithms to automatically generate parking lot section maps, significantly reducing complexity and cost
2Ease of manufacture
If open data sources like OSM are used, then cost is reduced, but coverage and maintenance quality deteriorate
Solution Approach 1:
The system continuously collects feedback from actual vehicle parking behaviors and uses this real-world data to automatically update and maintain parking lot section information, ensuring high reliability and current accuracy without relying on static open data sources
Solution Approach 2:
The patent performs preliminary actions by pre-processing vehicle GPS data to identify parking patterns and pre-defining the methodology for section determination, which enables continuous automatic maintenance and updates of parking lot section information without requiring manual intervention
3Extent of automation
If clustering based on feature amounts is performed, then automation is improved, but computational complexity increases
Solution Approach 1:
The patent segments the complex task of parking lot section determination into smaller manageable steps: dividing the map into grid cells, calculating parking feature amounts for each cell, and performing clustering only on relevant grid cells, which reduces overall computational complexity while maintaining high automation
Data Source
AI summary
A parking lot zone estimating device 30 comprises: a vehicle parking information recording unit 312 for determining that a vehicle is parked if there is no change in positional information for at least a predetermined period of time, and recording, as vehicle parking information, a parking location of the vehicle and a parking time that can be calculated from time information; a parking feature quantity calculating unit 313 for setting, as feature quantities of grids into which a region included in map information is divided, at least one of a parking frequency and a vehicle parking time, calculated on the basis of the vehicle parking information for each grid; and a facility parking lot estimating unit 314 for extracting the grids located within a prescribed range from a facility, and clustering the grids using the feature quantity to identify a parking lot of the facility.