Precipitation Estimation Using Adjacent-Area Vehicle Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing precipitation level estimation systems, such as those described in JP 2020-052962 A, face challenges in accurately estimating precipitation levels due to the wiper operation mode correlating less with precipitation intensity during intense rainfall, making it difficult to calculate the precipitation amount in a specific area.
Innovation Solution
A precipitation level estimation system utilizing machine learning to calculate precipitation amounts by acquiring detection data from vehicles, including features from past detection data in adjacent areas, and inputting these features into a precipitation amount estimation model to accurately estimate precipitation levels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If wiper operation mode data is used to estimate precipitation intensity, then the estimation method is simple and data acquisition is easy, but the estimation accuracy deteriorates during intense precipitation when wiper mode is fixed to high
Solution Approach 1:
The patent combines multiple data sources including wiper operation mode data, vehicle speed data, and weather information from multiple vehicles in a specific area. By merging these diverse data types, the system overcomes the limitation of wiper mode saturation during intense precipitation while maintaining the simplicity of using existing vehicle sensor data.
Solution Approach 2:
The patent introduces a server as an intermediary that collects, processes, and analyzes precipitation-related data from multiple vehicles. The server performs machine learning-based analysis to estimate precipitation levels, acting as a mediator between individual vehicle sensors and the final precipitation estimation, thereby improving accuracy through aggregated data while keeping individual vehicle systems simple.
2Device complexity
If data from a single area and time period is used, then the calculation is simple, but the precipitation level estimation accuracy is insufficient
Solution Approach 1:
The patent segments the data collection process by dividing the observation area into multiple regions and collecting data from different time periods. This segmentation allows the system to capture spatial and temporal variations in precipitation patterns, improving estimation accuracy while maintaining manageable calculation complexity through structured data organization.
Solution Approach 2:
The patent adds spatial and temporal dimensions to the data collection by gathering information from multiple vehicles across different locations and time periods. This multi-dimensional approach transforms the estimation problem from a single-point measurement to a comprehensive spatial-temporal analysis, significantly improving precipitation level accuracy.
Data Source
AI summary
A precipitation level estimation system includes a calculation unit configured to: acquire detection data related to precipitation; and calculate a precipitation amount in a predetermined area included in an observation target area. The calculation unit is configured to: calculate a first feature in a first predetermined period based on first detection data acquired from a first area including the predetermined area, the first predetermined period being a period that is past with respect to present; calculate a second feature in a second predetermined period in a second area adjacent to the first area based on second detection data acquired in the second area, the second predetermined period being a period that is past with respect to the first predetermined period; calculate the precipitation amount using a precipitation amount estimation model using the first feature and the second feature as variables; and calculate the precipitation amount based on the features.