Vehicle DTE Prediction Using Bayesian Driving Pattern Learning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional electric vehicles face inaccuracies in distance to empty (DTE) prediction due to differences between predicted values and actual mileage, especially when accounting for varying driving patterns, battery health changes, and vehicle friction, leading to difficulty in reflecting personalized driving conditions.
Innovation Solution
A DTE prediction apparatus and method using a linear regression model based on Bayesian probability distribution, which learns and updates models with new driving data, and generates clustering models to personalize driving patterns, correcting DTE predictions based on similar driving times and conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If conventional DTE calculation uses average driving characteristics of multiple arrays, then calculation simplicity is maintained, but prediction accuracy deteriorates due to inability to reflect personalized driving patterns
Solution Approach 1:
The patent changes the parameter basis from fixed average values to dynamically updated personalized parameters. The system collects actual driving data and updates driving characteristics parameters individually for each vehicle, transforming the static average-based calculation into a dynamic personalized calculation that adapts to each vehicle's unique driving patterns while maintaining computational feasibility.
2Measurement precision
If DTE model is updated frequently with new driving data, then prediction accuracy improves, but computational complexity and data processing burden increase
Solution Approach 1:
The patent applies preliminary action by collecting and storing driving data continuously in the background before actual DTE calculation is needed. The system pre-processes driving characteristics and maintains updated parameter sets ready for quick retrieval during DTE computation, avoiding the need for complex real-time processing when DTE is actually required.
Solution Approach 2:
The system implements self-service through automatic driving data collection and model updating without requiring manual intervention. The vehicle's existing sensors and control units continuously provide driving data, and the system automatically updates driving characteristics parameters, eliminating the need for external calibration or manual data entry while maintaining accurate personalized models.
3Stability of the object's composition
If conventional DTE calculation uses fixed average driving efficiency, then stability is maintained, but adaptability to changing driving conditions deteriorates
Solution Approach 1:
The patent transforms the static fixed average driving efficiency into a dynamic personalized driving characteristic that evolves over time. The system continuously updates driving parameters based on actual vehicle operation, allowing the DTE calculation to adapt to changing driving conditions, battery health, and vehicle friction while maintaining computational stability through systematic data collection and gradual parameter adjustment.
Data Source
AI summary
A vehicle distance to empty (DTE) prediction apparatus and a method therefor include a processor configured to define a DTE prediction model of a vehicle based on a linear regression model reflecting a Bayesian probability distribution, and to predict a DTE by learning the DTE prediction model based on driving data of the vehicle; and a storage configured to store data and algorithms driven by the processor.