… This technology is still in an early research stage and needs to be proven to work in complex
driving scenarios, eg intersections with more interacting traffic participants. Additionally …

Achieving excellent generalization on perceiving real traffic scenarios with diversity is the
long-term goal for building robust autonomous driving systems. In this paper, we propose to …

… rethink the open-world object detection in the context of autonomous driving scenarios, and
… More importantly, we call for attention to the more realistic setting for open-world perception …

… Regarding the extensive search space of establishing world models in real-world driving
scenarios, we introduce a two-stage training strategy for DriveDreamer. This strategy is …

… At each time step, we leverage the world model to generate predicted future scenarios for
trajectory candidates sampled from the planner, evaluate the futures using an image-based …

… To the best of our knowledge, this is the first study demonstrating the possibility of generating
realistic and continuous physical-world tests for practical autonomous driving scenarios. …

… In the real world: we test 10 models listed in Table I on 9 driving scenarios. In total we report
results … In the real world, for each scenario, we measure the percentage of distance driven …

… We are detailing the approach in the current work and providing two real-world cases to
demonstrate the implementation and result of this approach. The two cases involve two …

… This evolution is essential in the face of real-world scenarios’ complexity and unpredictability,
where traditional AI approaches struggle to replicate the depth and variability of human …

… We decided for this approach to allow the language to be system independent and to reflect
a real world driving scenario. In practice, a driving mission is given to the driver or ADS as a …