期刊
2022 IEEE/RSJ INTERNATIONAL CONFERENCE ON INTELLIGENT ROBOTS AND SYSTEMS (IROS)
卷 -, 期 -, 页码 12534-12541出版社
IEEE
DOI: 10.1109/IROS47612.2022.9981029
关键词
-
类别
资金
- NSF [1834701, 1839511, 1724341, 2038853]
- ONR [N00014-19-1-2496]
- Division Of Computer and Network Systems
- Direct For Computer & Info Scie & Enginr [2038853] Funding Source: National Science Foundation
This work proposes a behavior-aware Trajectory Autoencoder (TAE) model that explicitly models drivers' behavior to provide more information for trajectory generation and prediction, enhancing decision making in critical scenarios.
Trajectory generation and prediction are two interwoven tasks that play important roles in planner evaluation and decision making for intelligent vehicles. Most existing methods focus on one of the two and are optimized to directly output the final generated/predicted trajectories, which only contain limited information for critical scenario augmentation and safe planning. In this work, we propose a novel behavior-aware Trajectory Autoencoder (TAE) that explicitly models drivers' behavior such as aggressiveness and intention in the latent space, using semi-supervised adversarial autoencoder and domain knowledge in transportation. 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. Experimental results demonstrate that our method achieves promising performance on both trajectory generation and prediction.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据