Neural Network Vehicle Lateral Control Trajectory Tracking
Solution Overview
Problem
Vehicle lateral control systems face challenges in achieving accurate trajectory tracking and control smoothness due to complex vehicle and tire dynamics, especially at multiple speeds, where high-fidelity modeling is difficult to obtain.
Innovation Solution
A processor-implemented method using a feedforward artificial neural network (ANN) generates steering angle commands based on desired waypoint data and vehicle speed without modeling vehicle or tire dynamics, trained with data from various driving scenarios to achieve automatic vehicle lateral control.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If high-fidelity vehicle dynamics modeling is used to achieve accurate trajectory tracking, then control precision is improved, but device complexity and difficulty of obtaining the model increase
Solution Approach 1:
The patent replaces complex mechanical dynamics modeling with a neural network-based data-driven approach. Instead of using traditional vehicle dynamics models that require precise mathematical representations of tire forces, suspension, and vehicle mass distribution, the system uses a neural network trained on actual driving data to directly map steering inputs to trajectory outputs. This substitution eliminates the need for complex analytical models while maintaining or improving tracking accuracy.
Solution Approach 2:
The patent changes the fundamental parameters of the control system by transitioning from physics-based parameters (tire friction coefficients, mass, moment of inertia) to data-based parameters (neural network weights and biases learned from training data). This parameter transformation allows the system to adapt to varying vehicle conditions without requiring recalibration of physical parameters, thereby simplifying the overall system while maintaining precision.
2Manufacturing precision
If high-fidelity vehicle dynamics modeling is used to achieve accurate trajectory tracking, then control precision is improved, but ease of manufacture and implementation decrease
Solution Approach 1:
The patent replaces complex mechanical dynamics modeling with a neural network-based data-driven approach. Instead of using traditional vehicle dynamics models that require precise mathematical representations of tire forces, suspension, and vehicle mass distribution, the system uses a neural network trained on actual driving data to directly map steering inputs to trajectory outputs. This substitution eliminates the need for complex analytical models while maintaining or improving tracking accuracy.
Solution Approach 2:
The neural network model learns vehicle-specific characteristics automatically during the training phase by processing actual driving data collected from the specific vehicle. This self-learning capability eliminates the need for manual model parameter tuning, tire force calibration, or complex system identification procedures that would otherwise be required for high-fidelity dynamics modeling, thereby significantly easing implementation.
3Manufacturing precision
If complex vehicle and tire dynamics modeling is used, then trajectory tracking accuracy is improved, but control smoothness deteriorates
Solution Approach 1:
The patent replaces complex mechanical dynamics modeling with a neural network-based data-driven approach. Instead of using traditional vehicle dynamics models that require precise mathematical representations of tire forces, suspension, and vehicle mass distribution, the system uses a neural network trained on actual driving data to directly map steering inputs to trajectory outputs. This substitution eliminates the need for complex analytical models while maintaining or improving tracking accuracy.
Solution Approach 2:
The patent implements a dynamic control approach where the neural network continuously adapts its predictions based on current vehicle state and desired trajectory. The system uses a receding horizon control strategy where the neural network generates steering commands by considering future trajectory points, allowing smooth transitions and anticipatory control adjustments that maintain both accuracy and smoothness throughout the maneuver.
Data Source
AI summary
A processor-implemented method in a vehicle for performing automatic vehicle lateral control using a neural network for trajectory tracking is provided. The method includes receiving a desired vehicle trajectory; constructing, from the desired trajectory, desired waypoint data for a small number of desired waypoints at different look ahead distances in front of the center of gravity of the vehicle; and generating a steering angle command using a feedforward artificial neural network (ANN) as a function of vehicle speed and the desired waypoint data without modeling vehicle or tire dynamics. The vehicle steering actuator system provides steering control to cause vehicle steering to attempt to achieve the steering angle command.