Steering Comfort Evaluation in Autonomous Vehicle Take-Over Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional steering comfort evaluation methods are not suitable for autonomous vehicles as they are non-online and based on human driving scenarios, lacking real-time calculation capabilities and applicability in daily driving conditions.
Innovation Solution
An evaluation method and system for steering comfort in autonomous vehicles that uses vehicle data from sensors and a well-trained comfort model library for log probability matching, enabling online comfort evaluation during actual driving cycles and improving human-machine interaction safety.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional steering comfort evaluation methods are used, then subjective evaluation scores can be obtained, but real-time calculation capability is lost and applicability in daily driving conditions deteriorates
Solution Approach 1:
The patent replaces traditional mechanical evaluation methods (questionnaires, subjective scoring) with a computational model system. The comfort evaluation model uses vehicle state parameters and driver operation parameters as inputs to calculate comfort scores in real-time, substituting the mechanical process of subjective evaluation with an automated computational system that maintains both accuracy and real-time capability.
2Measurement precision
If traditional steering comfort evaluation methods are used, then subjective evaluation scores can be obtained, but applicability in autonomous vehicle scenarios deteriorates
Solution Approach 1:
The patent creates a dynamic evaluation system that adapts to different driving scenarios including autonomous vehicle operations. The model dynamically selects and processes relevant parameters based on the current operating context, allowing it to maintain applicability across both traditional human-driven and autonomous vehicle scenarios while preserving evaluation accuracy.
Solution Approach 2:
The comfort evaluation model is designed with universal applicability across multiple vehicle operating modes. It can evaluate steering comfort in both traditional human-driven scenarios and autonomous vehicle scenarios by processing the same types of input parameters (vehicle state, driver operation) regardless of the specific driving context, making it versatile and adaptable.
3Measurement precision
If detailed vehicle data collection is implemented, then evaluation accuracy is improved, but system complexity increases
Solution Approach 1:
The patent extracts only the essential parameters needed for comfort evaluation from the complete vehicle data set. Instead of processing all available vehicle data, the model selectively uses vehicle state parameters (speed, acceleration, steering angle) and driver operation parameters (steering torque, hand position), filtering out unnecessary information to maintain evaluation accuracy while reducing system complexity.
Data Source
AI summary
The present disclosure relates to an evaluation method and system for steering comfort in a human machine cooperative take-over control process of an autonomous vehicle, and a storage medium, thereby avoiding a defect that most comfort evaluation methods are applicable only in laboratory conditions, and also solving a problem that most comfort evaluation methods do not consider non-traditional control characteristics of a driver in a take-over control process. The method of the present disclosure includes: acquiring vehicle data information in a current driving cycle, and preprocessing the vehicle data information to form preprocessed vehicle data information; and using the preprocessed vehicle data information as an input to a well-trained comfort model library, and performing log probability matching calculation on the basis of the comfort model library, to form a comfort recognition result corresponding to comfort.