Wheel Torque Control Using RBF Networks Under Varying Road Friction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Vehicle dynamics control systems require extensive calibration for each vehicle type and road condition, complicating the control of wheel torque and stability, especially under varying friction and normal force conditions.
Innovation Solution
A method using a trained radial basis function network to determine torque changes based on input values such as wheel slip, acceleration, and historical force and torque data, allowing for robust and reproducible control of wheel torque.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional vehicle dynamics control systems are used with extensive calibration for each vehicle type and road condition, then control accuracy can be improved, but device complexity and calibration time increase significantly
Solution Approach 1:
The patent replaces traditional mechanical calibration procedures with a neural network-based system that automatically learns optimal control parameters. The neural network is trained offline using measured data from various road surfaces and vehicle conditions, then deployed to perform real-time torque control without requiring manual recalibration for each condition. This substitutes the complex calibration process with a data-driven computational approach.
Solution Approach 2:
The patent changes the control parameters from fixed calibration values to dynamically adjustable parameters generated by the neural network. The network takes inputs such as wheel slip, vehicle speed, and road surface characteristics, then outputs optimized torque values that adapt to changing conditions. This allows the system to maintain high control accuracy across different road surfaces without extensive manual calibration for each scenario.
2Adaptability or versatility
If traditional control systems are extensively calibrated for different road surface conditions, then adaptability to various surfaces can be improved, but loss of time during calibration and operation increases
Solution Approach 1:
The patent performs the calibration action in advance by training the neural network offline using comprehensive datasets that include multiple road surface conditions, vehicle types, and operating scenarios. This preliminary training phase captures the optimal control strategies for various conditions, allowing the system to immediately adapt to new situations during operation without requiring time-consuming recalibration. The network is pre-loaded with knowledge about different road surfaces and can generalize to unseen conditions.
3Productivity
If a trained radial basis function network is used to determine torque changes, then productivity and control speed can be improved, but manufacturing precision of the control system increases
Solution Approach 1:
The patent replaces traditional rule-based control algorithms with a neural network that learns optimal control strategies from data. The radial basis function network processes inputs and generates torque commands at high speed, enabling rapid response to changing driving conditions. The network architecture with radial basis functions provides smooth, continuous control outputs that are well-suited for torque regulation, achieving both high productivity and acceptable manufacturing precision through data-driven learning rather than manual tuning.
Data Source
AI summary
A method for controlling a torque of at least one wheel of a mobile platform. The method includes: providing at least one current slip value of the wheel and at least one current wheel acceleration of the wheel as input values; providing a trained radial basis function network designed to determine, by means of the input values, at least one torque change as an output value for control of the at least one wheel; and determining a current torque change, by means of the trained radial basis function network and the provided input values, for control of the torque.