Autonomous Driving (AD) faces crucial hurdles for commercial launch, notably in the form of diminished public trust and safety concerns from long-tail unforeseen driving scenarios.
In the context of safe exploration, Reinforcement Learning (RL) has long grappled with the challenges of balancing the tradeoff between maximizing rewards and minimizing safety violations, particularly in complex environments with contact-rich or non-smooth dynamics, and when dealing with high-dimensional pixel observations.
Trajectory generation and trajectory prediction are two critical tasks for autonomous vehicles, which generate various trajectories during development and predict the trajectories of surrounding vehicles during operation, respectively.
We conduct extensive experiments to demonstrate that our supervised method based on contrastive learning and unsupervised method based on reconstruction with semantic latent space can significantly improve the performance of anomalous trajectory detection in their corresponding settings over various baseline methods.
In this paper, we present a novel adversarial training method for trajectory prediction.
It is quite challenging to ensure the safety of reinforcement learning (RL) agents in an unknown and stochastic environment under hard constraints that require the system state not to reach certain specified unsafe regions.
Moreover, GROCET provides differentiable global robustness, which is leveraged in the training of globally robust neural networks.
In addition, experiments show that our method can significantly improve the system's robust generalization to unseen patterns of attacks.
Our model addresses trajectory generation and prediction in a unified architecture and benefits both tasks: the model can generate diverse, controllable and realistic trajectories to enhance planner optimization in safety-critical and long-tailed scenarios, and it can provide prediction of critical behavior in addition to the final trajectories for decision making.
The experiment results demonstrate that our approach can effectively mitigate the impact of adversarial attacks and can achieve 55% to 90% improvement over the original OpenPilot.