Vehicle Lateral Control Using Reference Steering Angle MPC
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle lateral control systems struggle to efficiently and reliably prevent side collisions and maintain lane centering, particularly in curves, due to inaccuracies in steering angle control.
Innovation Solution
A method utilizing a reference steering angle based on the product of the vehicle's wheel base, road curvature, understeer gradient, and lateral acceleration, integrated into a model predictive controller with a modified cost function to optimize steering wheel angle for lane centering and changes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional lateral control systems are used, then the system structure is simple, but the control accuracy and reliability are insufficient, especially in curves
Solution Approach 1:
The patent modifies the cost function parameters in the model predictive controller by incorporating a reference steering angle term that accounts for vehicle dynamics parameters (understeer gradient, lateral acceleration, wheel base, road curvature). This parameter enhancement improves control reliability without requiring additional hardware or complex system architecture.
Solution Approach 2:
The patent replaces conventional simple proportional-integral-derivative (PID) or linear quadratic regulator (LQR) control mechanisms with a model predictive control (MPC) framework that uses a modified cost function. This substitution improves control accuracy and reliability, particularly in curved road conditions, while maintaining computational efficiency through the structured cost function formulation.
2Measurement precision
If conventional steering control is used, then the system is easy to operate, but the lane centering precision and overshoot reduction are insufficient
Solution Approach 1:
The patent calculates a reference steering angle in advance using vehicle dynamics parameters and road curvature information before applying it in the cost function. This preliminary calculation of the optimal steering angle based on predicted vehicle state and road geometry enables precise lane centering while the MPC framework automatically handles the control execution, maintaining ease of operation.
Solution Approach 2:
The modified cost function incorporates feedback from vehicle state measurements (lateral position, heading angle, lateral acceleration) and road curvature information to continuously adjust the reference steering angle. This feedback mechanism improves lane centering precision by compensating for disturbances and model inaccuracies while the automated MPC implementation maintains ease of operation.
3Manufacturing precision
If simple cost function is used in model predictive controller, then the computation is fast, but the steering angle optimization and overshoot reduction are insufficient
Solution Approach 1:
The patent enhances the cost function by adding a reference steering angle term that incorporates vehicle dynamics parameters (understeer gradient, lateral acceleration, wheel base) and road curvature. This parameter enrichment improves steering angle precision and overshoot reduction while the analytical formulation of the reference steering angle keeps computational requirements manageable for real-time implementation.
Solution Approach 2:
The reference steering angle is calculated in advance using a closed-form expression based on vehicle dynamics parameters and road geometry. This preliminary calculation avoids iterative optimization during the control loop, maintaining fast computation speed while achieving precise steering angle optimization and reduced overshoot through the enhanced cost function.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A computer implemented method for lateral control of a vehicle comprises the following steps carried out by computer hardware components: determining a reference steering angle for the vehicle; determining at least one control variable based on a model predictive controller using the reference steering angle; and laterally controlling the vehicle based on the at least one control variable.