Mind the Noise: Sensitivity of Transformer-based Interaction-Aware Trajectory Prediction Models to Noisy Data

2026-06-19

AI Summary

This study addresses the significant performance degradation of Transformer-based trajectory prediction models under perception and localization noise encountered in real-world deployments. It presents the first systematic quantification of how varying levels of state noise affect state-of-the-art interaction-aware trajectory prediction models, revealing their pronounced vulnerability in realistic noisy conditions. The authors employ an attention-based Transformer architecture to model multi-agent interactions and evaluate its performance on state data injected with controlled noise. Experimental results demonstrate that even modest noise levels reduce prediction accuracy by a factor of 1.3, while more realistic high-noise scenarios incur accuracy losses up to 3.9×, underscoring the critical need for training and evaluation benchmarks that better reflect real-world sensor imperfections.

Abstract

Trajectory prediction allows autonomous vehicles to anticipate the future behavior of surrounding objects (or agents) and, accordingly, maximize the safety and efficiency of their driving. State-of-the-art Transformed-based interaction-aware trajectory prediction models, which rely on attention mechanisms to capture multi-agent interactions and maximize prediction accuracy, are commonly trained and evaluated on long-range high-quality datasets. These datasets are typically obtained by aggregating data from multiple vehicles or drones and removing any object detection or tracking noise offline. Yet, information about a surrounding object's state (its position, speed, heading) is far from being noiseless in real-world deployments. Object state estimation is affected by perception uncertainties and localization errors that can be particularly large for objects received via Vehicle-to-Everything (V2X) communications. In this paper, we analyze the impact of noisy object state information on the trajectory prediction accuracy of a state-of-the-art Transformer-based interaction-aware trajectory prediction model. Our study demonstrates that trajectory prediction accuracy can rapidly deteriorate as the noise intensity increases. Numerical results show that the prediction accuracy can reduce by a 1.3x factor under small noise levels and by as much as a 3.9x factor under the highest (yet realistic) noise conditions. These findings reveal the strong sensitivity of trajectory prediction models to noisy data, underscoring the need for more realistic training and evaluation datasets as well as noise mitigation strategies.

Problem

• trajectory prediction
• noisy data
• Transformer-based models
• V2X communications
• object state estimation

Innovation

• trajectory prediction
• Transformer-based models
• noise sensitivity
• V2X communication
• interaction-aware modeling