Trajectory prediction is a key element of autonomous vehicle systems, enabling them to anticipate and react to the movements of other road users. Robustness testing through adversarial methods is essential for evaluating the reliability of these prediction models. However, current approaches tend to focus solely on manipulating model inputs, which can generate unrealistic scenarios and overlook critical vulnerabilities. This limitation may result in incomplete assessments of model performance in real-world conditions. The specific effects of more comprehensive adversarial attacks on trajectory prediction models have not been thoroughly investigated. In this work, we demonstrate that by perturbing both model inputs and anticipated future states, we can uncover previously undetected weaknesses and provide a more realistic evaluation of model robustness. Our novel approach incorporates dynamical constraints and preserves tactical behaviors, enabling more effective and realistic adversarial attacks. We introduce new performance measures to assess the realism and impact of these adversarial trajectories. Testing our method on a state-of-the-art prediction model reveals significant increases in prediction errors and collision rates under adversarial conditions. Qualitative analysis further shows that our attacks can expose critical weaknesses, such as the model’s inability to detect potential collisions in what appear to be safe predictions. These results underscore the need for more comprehensive adversarial testing to better evaluate and improve the reliability of trajectory prediction models for autonomous vehicles. To support further research in this area, we provide an open-source framework for studying adversarial robustness in trajectory prediction. This work advances adversarial testing techniques, contributing to the safety and reliability of autonomous driving systems.

For trajectory prediction within autonomous vehicle planning and control, conditional variational autoencoders (CVAEs) have shown promise in accurate and diverse modeling of agent behaviors. Besides accuracy, explainability is also crucial for the safety and acceptance of learning-based autonomous systems, especially in autonomous driving. However, the latent distributions learned by CVAE models are often implicit and thus possess low explainability. To address this, we propose a semi-supervised generative modeling framework, \textbf{\textit{PrefCVAE}}, which utilizes partially and weakly labelled preference pairs to imbue the CVAE's latent representation with semantic meaning. This approach enables the system to estimate measurable attributes of the agents, and to generate manipulable predictions under the CVAE framework. Results show that incorporating our preference loss allows a CVAE-based model to make conditional predictions using the semantic factor of prediction average velocity. Our augmented framework also does not significantly degrade the baseline accuracy of prediction. Additionally, we show that the latent values learned with PrefCVAE better represent the semantic information contained in the data. Finally, we discuss the potential of this loss design to extend to other machine learning applications beyond trajectory prediction, as well as essential tricks for adaptation of human labeling. We hope that our empirical study offers the broader representation learning community a fresh perspective on inductive bias for disentangled and explainable latent representations in deep generative models. Specifically, we demonstrate that preference pair supervision, a simple and cost-effective approach, can effectively aid in learning semantic meanings for sampling-based generative models like the CVAE.