Vehicle Driving Load Prediction Error Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current automatic transmission control methods cannot accurately predict and compensate for driving loads based on road shapes ahead, leading to errors in speed change gear determination, especially when using commercial navigation systems with uncertain accuracy.
Innovation Solution
A method and system that select prediction positions to measure and correct driving conditions by calculating errors between predicted and real driving loads, using vehicle position and road information to adjust predictions without requiring a high-definition navigation system, thereby improving prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a high definition navigation system is mounted to reduce errors in speed change gear determination, then prediction accuracy of driving load is improved, but manufacturing cost increases greatly
Solution Approach 1:
The patent creates a virtual copy of the road profile by measuring actual vehicle acceleration and comparing it with predicted acceleration from navigation data. This virtual replication allows the system to learn and compensate for navigation system errors without requiring expensive high-definition navigation hardware, thereby achieving high prediction accuracy while maintaining low manufacturing cost
Solution Approach 2:
The system implements a feedback mechanism where the difference between actual vehicle acceleration (measured by accelerometers) and predicted acceleration (from navigation road profile data) is continuously calculated. This error signal is used to adjust and refine the driving load prediction, enabling the system to compensate for navigation inaccuracies and achieve high prediction accuracy using only commercial navigation systems
2Device complexity
If commercial navigation system is used to predict road shape, then device complexity is reduced, but measurement precision of driving load deteriorates
Solution Approach 1:
The patent replaces reliance on complex mechanical measurement systems with a software-based learning approach. Instead of using expensive high-definition navigation hardware, the system uses standard commercial navigation data combined with vehicle sensor data (accelerometers) and applies machine learning algorithms to compensate for errors, thereby reducing device complexity while maintaining measurement precision
Solution Approach 2:
The system performs self-calibration by using the vehicle's own acceleration measurements to identify and correct errors in the navigation system's road profile data. The vehicle essentially serves its own calibration needs by comparing predicted versus actual acceleration, eliminating the need for external high-precision measurement equipment
Data Source
AI summary
A method of predicting a driving condition of a vehicle may include selecting a first prediction position where a vehicle is predicted to pass afterward while driving and predicting a first driving load of the vehicle at the first prediction position; when the vehicle reaches the first prediction position, measuring a driving condition of the vehicle at the first prediction position; and predicting a second driving load at a second prediction position where the vehicle is predicted to pass afterward by reflecting an error between the first driving load at the first prediction position and the real driving condition at the first prediction position.