Constraint-Aware Diffusion Models for Trajectory Optimization
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2024
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866929371083702272 |
|---|---|
| author | Li, Anjian Ding, Zihan Dieng, Adji Bousso Beeson, Ryne |
| author_facet | Li, Anjian Ding, Zihan Dieng, Adji Bousso Beeson, Ryne |
| contents | The diffusion model has shown success in generating high-quality and diverse solutions to trajectory optimization problems. However, diffusion models with neural networks inevitably make prediction errors, which leads to constraint violations such as unmet goals or collisions. This paper presents a novel constraint-aware diffusion model for trajectory optimization. We introduce a novel hybrid loss function for training that minimizes the constraint violation of diffusion samples compared to the groundtruth while recovering the original data distribution. Our model is demonstrated on tabletop manipulation and two-car reach-avoid problems, outperforming traditional diffusion models in minimizing constraint violations while generating samples close to locally optimal solutions. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_00990 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Constraint-Aware Diffusion Models for Trajectory Optimization Li, Anjian Ding, Zihan Dieng, Adji Bousso Beeson, Ryne Machine Learning Robotics The diffusion model has shown success in generating high-quality and diverse solutions to trajectory optimization problems. However, diffusion models with neural networks inevitably make prediction errors, which leads to constraint violations such as unmet goals or collisions. This paper presents a novel constraint-aware diffusion model for trajectory optimization. We introduce a novel hybrid loss function for training that minimizes the constraint violation of diffusion samples compared to the groundtruth while recovering the original data distribution. Our model is demonstrated on tabletop manipulation and two-car reach-avoid problems, outperforming traditional diffusion models in minimizing constraint violations while generating samples close to locally optimal solutions. |
| title | Constraint-Aware Diffusion Models for Trajectory Optimization |
| topic | Machine Learning Robotics |
| url | https://arxiv.org/abs/2406.00990 |