Driving Pattern User Classification With Privacy-Preserving Features
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The classification of user types for optimizing battery performance in new energy vehicles is complex and poses a risk to information security due to the need to analyze sensitive driving data.
Innovation Solution
A user type identification method that obtains driving data including driving time, mileage, and speed, and analyzes daily driving duration, mileage, and frequency to categorize users without accessing sensitive information, using a preset identification model to classify users into types such as daytime and night online car-hailing, commuting private vehicle, and commercial vehicle types.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sensitive driving data such as driving routes are obtained and analyzed for user type classification, then user type identification accuracy is improved, but user information security deteriorates
Solution Approach 1:
The patent extracts only the necessary non-sensitive features (driving duration, mileage, frequency) from the complete driving data set, separating useful classification information from sensitive personal information. This extraction principle resolves the contradiction by obtaining sufficient data for accurate user type identification while excluding sensitive information that would compromise user privacy and security.
Solution Approach 2:
The patent introduces an intermediary processing layer (the preset identification model) that transforms raw driving data into aggregated statistical features before classification. This intermediary approach maintains identification accuracy by preserving essential driving patterns while eliminating sensitive details, thus resolving the contradiction between accuracy and security.
2Adaptability or versatility
If complex processing procedures are used for user type classification, then classification comprehensiveness is improved, but system complexity deteriorates
Solution Approach 1:
The patent segments the user type classification process into distinct modules: data acquisition, feature extraction (driving duration, mileage, frequency), model analysis, and classification output. This segmentation maintains comprehensive classification capability while reducing overall system complexity by organizing the processing procedure into manageable, independent stages.
Solution Approach 2:
The patent changes the parameters from detailed sensitive data (routes, locations) to aggregated statistical parameters (duration, mileage, frequency). This parameter transformation maintains the comprehensiveness of classification by preserving essential driving behavior patterns while simplifying the processing procedure and reducing data complexity.
Data Source
AI summary
Disclosed are a user type identification method, an electronic device, and a readable storage medium. The method includes: obtaining driving data in a preset time period, where the driving data includes at least a driving time, an accumulated driving mileage, and a driving speed in a vehicle driving process; obtaining to-be-analyzed data based on the driving data in the preset time period, where the to-be-analyzed data includes daily driving duration, a daily driving mileage, and a quantity of times of driving at each moment daily; and performing analysis on the to-be-analyzed data through a preset identification model, to obtain a user type.