arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2610.12910cs.RO
收藏

基于世界模型的采样式运动-动力学规划

Sampling-based Kinodynamic Planning with World Models

Purdue University(普渡大学)

Yitian Gao, Hanlan Yang, Zachary Kingston

AI总结:

本文提出两种适用于世界模型隐空间的采样式运动-动力学规划策略,在模拟和真实推方块任务中,其规划效率和成功率均优于CEM、MPPI等优化器。

摘要:

视觉世界模型可从图像预测动作的后果,但利用其进行规划通常依赖CEM、MPPI等优化器。然而,这些优化器在长 horizon 任务中会陷入停滞,此类任务中进展需要目标未奖励的运动,例如推动方块前绕其移动。本文重新研究针对世界模型的采样式运动-动力学规划,将学习到的模型用作隐空间树搜索的传播器。核心见解是,为构型空间高维搜索设计的采样策略可直接应用于世界模型的隐空间,能快速扩展近视优化器从未采样的中间状态。我们提出两种高维隐空间探索策略:近邻引导RRT和基于投影的KPIECE,采用GPU原生批量扩展以实现高效设备利用。我们在四个模拟域和一个真实世界推方块任务中使用现成的DINO-WM进行评估。在模拟中,我们的方法达到或超过CEM和MPPI的成功率,规划时间最多降低13.8倍;在真实机器人上,成功率提升20%。

英文摘要:

Visual world models predict the consequences of actions from images, but planning with them is typically done by optimizers such as CEM and MPPI. However, these optimizers stall on long-horizon tasks where progress requires motion the objective does not reward, such as moving around a block before pushing it. This paper revisits kinodynamic sampling-based motion planning for world models, using the learned model as the propagator of a tree search in latent space. Our key insight is that the sampling strategies devised for high-dimensional search in configuration space apply directly to the latent space of a world model, where they rapidly expand through intermediate states that myopic optimizers never sample. We present two exploration strategies for the high-dimensional latent space, nearest-neighbor-guided RRT and projection-based KPIECE, with GPU-native batched expansion for efficient device utilization. We evaluate with an off-the-shelf DINO-WM on four simulated domains and a real-world block-pushing task. Our approaches match or exceed the success rate of CEM and MPPI at up to 13.8x lower planning time in simulation, and improve the success rate by 20% on the real robot.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑