Driving Risk Model Calibration Using Pedal-Based Risk Curves
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current driving risk identification systems for smart vehicles face challenges in adapting to the changing physiological and psychological states of drivers and individual differences, leading to reduced acceptance and safety, especially in complex traffic environments.
Innovation Solution
A method for calibrating a driving risk identification model by conducting free driving tests with an information acquisition device on a test vehicle, extracting key moments of pedal operations, defining risk level values, and using curve fitting to create a risk identification curve, which is then used to adjust the model's parameters for better adaptability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the driving risk identification model uses fixed parameters, then the model structure is simple, but it cannot adapt to changing driver states and individual differences, reducing system acceptance and safety
Solution Approach 1:
The patent transforms the static driving risk identification model into a dynamic one by introducing real-time parameter calibration mechanisms. The system continuously adjusts model parameters based on actual driving data, driver physiological states, and environmental conditions, enabling the model to adapt to changing driver states and individual differences while maintaining a relatively simple base structure.
Solution Approach 2:
The patent implements parameter changes by calibrating model parameters through curve fitting against actual driving behavior data. Key parameters such as risk threshold values, response time constants, and sensitivity coefficients are adjusted based on empirical data from multiple drivers, allowing the model to accommodate individual differences and varying driver states without fundamentally changing the model architecture.
2Reliability
If the driving risk identification system provides frequent warnings, then driving safety is improved, but driver acceptance decreases due to disturbances in normal driving
Solution Approach 1:
The patent applies local quality by differentiating warning strategies based on specific driving contexts and driver states. Instead of uniform frequent warnings, the system adjusts warning frequency and intensity according to the actual risk level, driver attentiveness, and environmental conditions. This allows critical safety warnings to be prominent while reducing unnecessary disturbances during normal driving situations.
Solution Approach 2:
The patent implements feedback mechanisms where the system monitors driver responses to warnings and adjusts subsequent warning behavior accordingly. By tracking whether drivers respond appropriately to warnings or exhibit signs of alarm fatigue, the system dynamically adjusts warning frequency and intensity, maintaining high safety standards while preserving driver acceptance through adaptive feedback control.
3Adaptability or versatility
If the driving risk identification model uses calibrated parameters based on actual driving data, then system acceptance and safety improve, but the calibration process requires extensive testing and data collection
Solution Approach 1:
The patent applies preliminary action by conducting extensive parameter calibration during the offline development phase using comprehensive driving data from multiple drivers and scenarios. The calibration process, including data collection, curve fitting, and parameter optimization, is performed beforehand to establish an initial set of calibrated parameters. This preliminary calibration reduces the need for extensive real-time adjustment during actual deployment, thereby improving model adaptability while minimizing calibration time loss in operational settings.
Data Source
AI summary
A method for calibrating a driving risk identification model includes: S1. establishing a vehicle platform by installing an information acquisition device on a test vehicle; S2. acquiring synchronized test data related to the test vehicle and driving scenarios; S3. extracting moments when the drivers start to press the accelerator pedal, start to release the accelerator pedal, start to press the brake pedal in the driving scenarios, and start to release the brake pedal, so as to define risk level values respectively corresponding to the moments; S4. obtaining a risk identification curve of the drivers in different driving scenarios, wherein the risk identification curve represents the drivers' judgment of the risk level over time; S5. using the risk identification curve to calibrate the driving risk identification model.