Driving Behavior Prediction Using Merged Template Weighting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing driving behavior prediction systems face challenges in accurately predicting driver behavior due to intra-individual differences in driving data, leading to reduced prediction accuracy and reliability.
Innovation Solution
A method and apparatus that acquire position and route information, determine distances to object points, and use driving data to predict behavior based on templates and weighting factors, updating these factors to reflect actual driving behavior and improve prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If model learning is conducted based only on driving data showing driving behaviors of the driver of the own vehicle, then the system can be simplified and data collection becomes easier, but the prediction accuracy deteriorates due to intra-individual differences in driving data
Solution Approach 1:
The patent merges multiple data sources including driving data from the own vehicle, other vehicles, and simulation results to create a comprehensive training dataset. This combination allows the model to learn from diverse driving patterns while accounting for intra-individual differences, thereby improving prediction accuracy without requiring an overly complex system architecture.
Solution Approach 2:
The trained behavior recognition model is designed to be universally applicable across multiple vehicles and drivers. By training on aggregated data from various sources, a single model can accurately predict driving behaviors for different individuals, eliminating the need for separate models for each vehicle-driver combination and simplifying the overall system.
2Adaptability or versatility
If models are generated based on plural driving data to make the apparatus compatible with various drivers, then adaptability improves, but the system complexity increases and requires extensive data collection infrastructure
Solution Approach 1:
The patent creates a universal behavior recognition model that can adapt to various drivers through training on aggregated driving data from multiple sources. This single multi-functional model replaces the need for multiple driver-specific models, achieving compatibility with various drivers while maintaining manageable system complexity.
Solution Approach 2:
The system performs preliminary training using aggregated driving data from multiple vehicles and simulation environments before deployment. This pre-training phase establishes a robust baseline model that can then be fine-tuned or directly applied to individual vehicles, reducing the complexity of real-time adaptation and enabling versatile driver compatibility.
3Measurement precision
If individual driver-specific models are generated to accurately predict each driver's behavior, then prediction accuracy improves, but the data collection and processing requirements increase significantly
Solution Approach 1:
The patent combines driving data from multiple vehicles, drivers, and simulation environments into a unified training dataset. This merging approach allows the system to learn common driving behavior patterns across individuals while capturing individual variations, achieving accurate predictions without requiring exhaustive data collection for each specific driver.
Solution Approach 2:
The system uses simulation environments to generate synthetic driving data that complements real-world data. This self-service approach to data generation reduces the burden of collecting extensive real driving data while still providing sufficient training material for accurate individual behavior prediction.
Data Source
AI summary
A driving behavior prediction apparatus includes, for accurately predicting a driving behavior, a position calculation unit for a subject vehicle position calculation, a route setting unit for setting a navigation route, a distance calculation unit for calculating a distance to a nearest object point, a parameter storage unit for storing a template weighting factor that reflects a driving operation tendency of a driver, a driving behavior prediction unit for predicting a driver's behavior based on vehicle information and the template weighting factor, a driving behavior recognition unit for recognizing driver's behavior at the object point, and a driving behavior learning unit for updating the template weighting factor so as to study the driving operation tendency of the driver in a case that a prediction result by the driving behavior prediction unit agrees with a recognition result by the driving behavior recognition unit.