AIBearisharXiv – CS AI · 8h ago7/10
🧠
Mind the Noise: Sensitivity of Transformer-based Interaction-Aware Trajectory Prediction Models to Noisy Data
Researchers demonstrate that transformer-based trajectory prediction models used in autonomous vehicles experience severe accuracy degradation when exposed to noisy real-world sensor data, with prediction accuracy declining by up to 3.9x under realistic noise conditions. The findings highlight a critical gap between idealized training environments and actual deployment scenarios, signaling the need for robust noise mitigation strategies in autonomous vehicle systems.