The effects of four-wheel steering on the path-tracking control of automated vehicles

Sungjin Lim,Illés Vörös,Yongseob Lim,Gábor Orosz

2024-12-04

Abstract:In this study, we analyze the stability of a path-tracking controller designed for a four-wheel steering vehicle, incorporating the effects of the reference path curvature. By employing a simplified kinematic model of the vehicle with steerable front and rear wheels, we derive analytical expressions for the stability regions and optimal control gains specific to different four-wheel steering strategies. To simplify our calculations, we keep the rear steering angle $\delta_r$ proportional to the front steering angle $\delta_f$ by using the constant parameter $a$, i.e., $\delta_r = a\delta_f$, where $\delta_f$ is calculated from a control law having both feedforward and feedback terms. Our findings, supported by stability charts and numerical simulations, indicate that for high velocities and paths of small curvatures, the appropriately tuned four-wheel steering controller significantly reduces lateral acceleration and enhances path-tracking performance when compared to using only front-wheel steering. Furthermore, for low velocities and large curvatures, the using negative a values (i.e., steering the rear wheels in the opposite direction than the front wheels) allows for a reduced turning radius, increasing the vehicle's capability to perform sharp turns in confined spaces like in parking lots or on narrow roads.

Systems and Control

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the stability and performance optimization in the path - tracking control of four - wheel - steering (4WS) vehicles. Specifically, the main research objectives include:

1. **Analyze stability**: Analyze the stability of the path - tracking controller designed for four - wheel - steering vehicles by introducing the influence of the reference path curvature. In the study, a simplified vehicle kinematic model was used, considering the situation where both the front wheels and the rear wheels can be steered, and the analytical expressions of the stable regions and the optimal control gains under different four - wheel - steering strategies were derived.

2. **Optimize control performance**: Explore how to improve the path - tracking performance of autonomous vehicles under different speeds and path conditions by appropriately adjusting the four - wheel - steering controller. In particular, under high - speed and small - curvature path conditions, an appropriately tuned four - wheel - steering controller can significantly reduce the lateral acceleration and enhance the path - tracking performance; while under low - speed and large - curvature path conditions, using a negative rear - wheel - steering angle (that is, the rear wheels turn in the opposite direction to the front wheels) can reduce the turning radius, enabling the vehicle to turn more flexibly in a narrow space (such as a parking lot or a narrow road).

3. **Verify the effectiveness of the method**: Verify the effectiveness of the proposed four - wheel - steering controller through stability charts and numerical simulation experiments. The research results show that under specific conditions, the four - wheel - steering system has obvious advantages compared with the front - wheel - steering - only system and can better meet the safety and handling performance requirements of autonomous vehicles.

### Formula summary

- Relationship between the rear - wheel - steering angle and the front - wheel - steering angle:
\[
\delta_r = a\delta_f
\]
where \(a\) is a constant parameter.

- Key equations in the vehicle kinematic model:
\[
\dot{x}_R = V\cos(\psi+\delta_r)
\]
\[
\dot{y}_R = V\sin(\psi+\delta_r)
\]
\[
\dot{\psi}=\frac{V\sin(\delta_f - \delta_r)}{f\cos\delta_f}
\]

- Feedback control laws in the controller design:
\[
\delta_f^{FB}=-k_1y_R - k_2\psi
\]
\[
\delta_r^{FB}=-k_3y_R - k_4\psi
\]
where \(k_3 = ak_1\) and \(k_4 = ak_2\).

- Characteristic equation in the stability analysis:
\[
D(\lambda)=\lambda^2+\frac{V}{f}(fak_1+(1 - a)k_2)\lambda+\frac{V^2}{f}((1 - a)k_1+(1 - ak_2)f\kappa_C^2)=0
\]

Through these formulas and analyses, the paper shows the potential of four - wheel - steering technology in improving the path - tracking performance of autonomous vehicles and provides theoretical basis and technical support for practical applications.