RETO: A Rotary-Enhanced Transformer Operator for High-Fidelity Prediction of Automotive Aerodynamics

2026-04-30

AI Summary

Existing neural operators struggle to accurately capture the complex spatiotemporal correlations in automotive aerodynamic fields, limiting high-fidelity prediction accuracy. This work proposes a novel neural operator, RETO, which innovatively integrates sinusoidal global positional encoding with rotation-based rotary positional encoding (RoPE). This fusion enhances local gradient resolution and relative displacement modeling while preserving translation invariance. Built upon a Transformer architecture and guided by information entropy analysis, RETO significantly outperforms current methods on the ShapeNet and DrivAerML datasets, achieving 16%–23% lower relative L2 errors in surface pressure and velocity field predictions. Moreover, its reduced information entropy peaks demonstrate superior capability in focusing on and preserving local features.

Abstract

Rapid aerodynamic evaluation is crucial for modern vehicle design, yet existing neural operators struggle to capture intricate spatial correlations. We propose the rotary-enhanced transformer operator (RETO), a novel neural solver featuring a dual-stage spatial awareness mechanism: sinusoidal-cosine encodings for global referencing and rotary positional encodings (RoPE) for relative displacements. RoPE encodes spatial relations via unitary rotations, enforcing translation invariance and enhancing local gradient resolution. RETO is validated on ShapeNet and the high-fidelity DrivAerML benchmark. On ShapeNet, RETO achieves a relative $L_2$ error of 0.063, outperforming RegDGCNN at 0.125 and representing a 16\% improvement over the Transolver baseline, which yields an error of 0.075. These performance gains are further amplified on the DrivAerML dataset, where RETO achieves relative $L_2$ errors of 0.089 for surface pressure and 0.097 for velocity. In comparison, Transolver results in errors of 0.116 and 0.121 for the same metrics, indicating that RETO achieves precision enhancements of 23\% and 19\%, respectively. For comprehensive comparison, the surface pressure and velocity errors for AB-UBT are 0.102 and 0.124, while RegDGCNN yields 0.235 and 0.312, respectively. Information-theoretical analysis shows that the entropy peak of RETO at 0.35 is significantly lower than that of Transolver at 0.75 under $10^4$ resolution, indicating a focused attentional mechanism capable of preserving localized gradients against global diffusion.

Problem

• automotive aerodynamics
• neural operators
• spatial correlations
• high-fidelity prediction
• rotary positional encodings

Innovation

• Rotary Positional Encoding
• Neural Operator
• Spatial Awareness Mechanism
• Automotive Aerodynamics
• High-Fidelity Prediction