Vehicle Dynamics Control for Simultaneous Optimization of Tire Emissions and Performance in EVs

TL;DR

This work tackles tire wear emissions from electric vehicles by introducing a dual-profile tire strategy: front axles use low-traction hard tires and rear axles use high-traction soft tires. A control framework distributes torque and applies steering corrections to replicate the performance of a fully soft-tire baseline while minimizing tire wear particles, using a bicycle-model dynamic foundation and a path-discretized optimal-control formulation. Across braking, straight, and curved trajectories, the approach achieves substantial tire-emission reductions (typically 45–60%) with preserved drivability, demonstrating a practical method to mitigate environmental impacts without compromising user experience. The method proves robust to variations in tire wear models and points to safe reinforcement learning as a promising avenue for adaptive, data-driven tire-emission control.

Abstract

In recent years, Electric Vehicles (EVs) have seen widespread public adoption. While EVs produce zero tailpipe emissions, they contribute to an increase in another type of vehicular emission: tire emissions. Battery-operated EVs are generally heavier than their combustion-engine counterparts and require greater acceleration forces, which their high-torque electric motors provide. This combination of increased weight and traction forces leads to higher tire emissions, which possess various adverse health and environmental effects. Here, we propose a control solution with promising results in mitigating tire wear in all-wheel-drive EVs. The idea is to utilize different tire profiles on each drive axis: a low-wear, low-traction axis and a high-wear, high-traction axis. Derived from detailed mathematical analyses, we propose a simple control scheme to counteract the performance difference from using the low-traction tires. The proposed control mechanism then distributes torque optimally between the two axes, maximizing usage from the low-wear axis and simultaneously maintaining stability and performance by leveraging high-traction tires. Through detailed numerical simulations, we demonstrate that the developed model significantly reduces tire emissions and maintains vehicle drivability and performance.

Figures (12)

• Figure 1: Bicycle vehicle model. Vehicle model with two-wheel axis, with steerable front wheel. The distance between the vehicle's center of gravity(CoG) to its front and rear axis are denoted $l_f$ and $l_r$, respectively. The vehicle is under five forces (colored in blue): Longitudinal and lateral forces from the front wheel $F_{v,f},\,F_{u,f}$, longitudinal and lateral forces from the rear wheel $F_{v,r},\,F_{u,r}$, as well as the air drag forces $F_d = C_dv^2$. The vehicle's longitudinal and lateral velocity are denoted $v$ and $u$, colored in red. The vehicle also has an angular velocity around its vertical $z$-axis, denoted $\delta$ and colored in yellow. This angular velocity along with the lateral and longitudinal velocities also created centripetal acceleration $a_{v,c}$ and $a_{u,c}$, colored in green.
• Figure 2: Curvilinear coordinate system. The figure shows the vehicle frame coordinates (in blue) and curvilinear frame coordinates (in red) in relation to one another.
• Figure 3: Flow chart of the control algorithm. The driver's steering input and the steering adjustment angle are denoted by $\sigma$ and $\Delta\sigma$, respectively. The vehicle movement is described by longitudinal velocity ($u$), lateral velocity ($v$), and yaw rate ($\delta$). The reference force signals for the front and rear tires are denoted by $\Tilde{F}_{v,f}$ and $\Tilde{F}_{v,r}$, respectively. The input torques to the front and rear wheels are denoted by $T_f$ and $T_r$, respectively.
• Figure 4: Traction versus treadwear curve. Approximated friction coefficient on dry asphalt against UTQG treadwear rating based on data from fifteen tires. Blue points are the collected data, the dashed line is the fit in Eq. \ref{['eq:w-vs-mu']}, and the shaded area marks the $90\%$ confidence interval (CI) for the fit.
• Figure 5: Friction coefficient versus slip curves. Longitudinal (left) and lateral (right) friction coefficient of the soft and hard tires used for simulation for the soft (red) tire and the hard (green) tire. The dotted line is the slip limits that the controller would impose on the tires