Enhancing Vehicle Aerodynamics with Deep Reinforcement Learning in Voxelised Models

TL;DR

This work tackles aerodynamic design optimisation by applying deep reinforcement learning to voxelised vehicle models. A PPO-based DRL agent learns to adjust voxel heights in a Unity-enabled, wind-tunnel-like environment to minimise drag and optimise related metrics such as kinetic energy and voxel collisions. Across three car models, the method achieves significant improvements in aerodynamic performance, with drag reductions and energy-flow improvements demonstrating the potential of DRL to navigate complex, high-dimensional design spaces. While promising for efficiency and sustainability, the study notes the absence of structural, manufacturability, and thermal considerations, indicating avenues for future work to integrate practical engineering constraints.

Abstract

Aerodynamic design optimisation plays a crucial role in improving the performance and efficiency of automotive vehicles. This paper presents a novel approach for aerodynamic optimisation in car design using deep reinforcement learning (DRL). Traditional optimisation methods often face challenges in handling the complexity of the design space and capturing non-linear relationships between design parameters and aerodynamic performance metrics. This study addresses these challenges by employing DRL to learn optimal aerodynamic design strategies in a voxelised model representation. The proposed approach utilises voxelised models to discretise the vehicle geometry into a grid of voxels, allowing for a detailed representation of the aerodynamic flow field. The Proximal Policy Optimisation (PPO) algorithm is then employed to train a DRL agent to optimise the design parameters of the vehicle with respect to drag force, kinetic energy, and voxel collision count. Experimental results demonstrate the effectiveness and efficiency of the proposed approach in achieving significant results in aerodynamic performance. The findings highlight the potential of DRL techniques for addressing complex aerodynamic design optimisation problems in automotive engineering, with implications for improving vehicle performance, fuel efficiency, and environmental sustainability.

Figures (9)

• Figure 1: Interaction of an agent with the environment in reinforcement learning.
• Figure 2: Original car prototypes.
• Figure 3: Generated height maps.
• Figure 4: Voxelised models.
• Figure 5: Aerodynamic simulation environment.