A5040264128- Autonomous Vehicle Technology and Safety
- Traffic Prediction and Management Techniques
- Mechanical Engineering and Vibrations Research
- Vehicle Dynamics and Control Systems
- Advanced Neural Network Applications
- Electric and Hybrid Vehicle Technologies
- Robotic Path Planning Algorithms
- Human Pose and Action Recognition
- Time Series Analysis and Forecasting
- Semantic Web and Ontologies
- AI-based Problem Solving and Planning
- Multimodal Machine Learning Applications
- Vehicle Noise and Vibration Control
- Human-Automation Interaction and Safety
- Real-time simulation and control systems
- Anomaly Detection Techniques and Applications
- Simulation Techniques and Applications
TH Bingen University of Applied Sciences
2024
Robert Bosch (Germany)
2023-2024
The release of nuPlan marks a new era in vehicle motion planning research, offering the first large-scale real-world dataset and evaluation schemes requiring both precise short-term long-horizon ego-forecasting. Existing systems struggle to simultaneously meet requirements. Indeed, we find that these tasks are fundamentally misaligned should be addressed independently. We further assess current state closed-loop field, revealing limitations learning-based methods complex scenarios value...
The field of motion prediction for automated driving has seen tremendous progress recently, bearing ever-more mighty neural network architectures. Leveraging these powerful models bears great potential the closely related planning task. In this work, we show that state-of-the-art can be converted into goal-directed planners. To end, propose a novel goal-conditioning method. Our key insight is conditioning on navigation goal at behaviour level outperforms other widely adopted methods, with...
Real-world autonomous driving systems must make safe decisions in the face of rare and diverse traffic scenarios. Current state-of-the-art planners are mostly evaluated on real-world datasets like nuScenes (open-loop) or nuPlan (closed-loop). In particular, seems to be an expressive evaluation method since it is based data closed-loop, yet covers basic This makes difficult judge a planner's capabilities generalize rarely-seen situations. Therefore, we propose novel closed-loop benchmark...
The CVAE is one of the most widely-used models in trajectory prediction for AD. It captures interplay between a driving context and its ground-truth future into probabilistic latent space uses it to produce predictions. In this paper, we challenge key components CVAE. We leverage recent advances VAE, foundation CVAE, which show that simple change sampling procedure can greatly benefit performance. find unscented sampling, draws samples from any learned distribution deterministic manner,...
Benchmarking vision-based driving policies is challenging. On one hand, open-loop evaluation with real data easy, but these results do not reflect closed-loop performance. the other, possible in simulation, hard to scale due its significant computational demands. Further, simulators available today exhibit a large domain gap data. This has resulted an inability draw clear conclusions from rapidly growing body of research on end-to-end autonomous driving. In this paper, we present NAVSIM,...
The field of motion prediction for automated driving has seen tremendous progress recently, bearing ever-more mighty neural network architectures. Leveraging these powerful models bears great potential the closely related planning task. In this letter we propose a novel goal-conditioning method and show its to transform state-of-the-art model into goal-directed planner. Our key insight is that conditioning on navigation goal at behaviour level outperforms other widely adopted methods, with...
Predicting the future motion of observed vehicles is a crucial enabler for safe autonomous driving. The field prediction has seen large progress recently with State-of-the-Art (SotA) models achieving impressive results on large-scale public benchmarks. However, recent work revealed that learning-based methods are prone to predict off-road trajectories in challenging scenarios. These can be created by perturbing existing scenarios additional turns front target vehicle while history left...