Latent Imagination Facilitates Zero-Shot Transfer in Autonomous Racing

被引:10
|
作者
Brunnbauer, Axel [1 ]
Berducci, Luigi [1 ]
Brandstatter, Andreas [1 ]
Lechner, Mathias [2 ]
Hasani, Ramin [3 ]
Rus, Daniela [3 ]
Grosu, Radu [1 ]
机构
[1] Tech Univ Wien TU Wien, CPS, Vienna, Austria
[2] Inst Sci & Technol Austria IST Austria, Klosterneuburg, Austria
[3] Massachusetts Inst Technol MIT, CSAIL, Cambridge, MA USA
基金
奥地利科学基金会; 欧洲研究理事会;
关键词
D O I
10.1109/ICRA46639.2022.9811650
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
World models learn behaviors in a latent imagination space to enhance the sample-efficiency of deep reinforcement learning (RL) algorithms. While learning world models for high-dimensional observations (e.g., pixel inputs) has become practicable on standard RL benchmarks and some games, their effectiveness in real-world robotics applications has not been explored. In this paper, we investigate how such agents generalize to real-world autonomous vehicle control tasks, where advanced model-free deep RL algorithms fail. In particular, we set up a series of time-lap tasks for an F1TENTH racing robot, equipped with a high-dimensional LiDAR sensor, on a set of test tracks with a gradual increase in their complexity. In this continuous-control setting, we show that model-based agents capable of learning in imagination substantially outperform model-free agents with respect to performance, sample efficiency, successful task completion, and generalization. Moreover, we show that the generalization ability of model-based agents strongly depends on the choice of their observation model. We provide extensive empirical evidence for the effectiveness of world models provided with long enough memory horizons in sim2real tasks.
引用
收藏
页码:7513 / 7520
页数:8
相关论文
共 50 条
  • [1] Zero-Shot Policy Transfer in Autonomous Racing: Reinforcement Learning vs Imitation Learning
    Hamilton, Nathaniel
    Musau, Patrick
    Lopez, Diego Manzanas
    Johnson, Taylor T.
    [J]. 2022 IEEE INTERNATIONAL CONFERENCE ON ASSURED AUTONOMY (ICAA 2022), 2022, : 11 - 20
  • [2] PencilNet: Zero-Shot Sim-to-Real Transfer Learning for Robust Gate Perception in Autonomous Drone Racing
    Pham, Huy Xuan
    Sarabakha, Andriy
    Odnoshyvkin, Mykola
    Kayacan, Erdal
    [J]. IEEE ROBOTICS AND AUTOMATION LETTERS, 2022, 7 (04): : 11847 - 11854
  • [3] Latent Embeddings for Zero-shot Classification
    Xian, Yongqin
    Akata, Zeynep
    Sharma, Gaurav
    Nguyen, Quynh
    Hein, Matthias
    Schiele, Bernt
    [J]. 2016 IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2016, : 69 - 77
  • [4] Zero-Shot Task Transfer
    Pal, Arghya
    Balasubramanian, Vineeth N.
    [J]. 2019 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2019), 2019, : 2184 - 2193
  • [5] Salient Latent Features For Zero-shot Learning
    Pan, Zongrong
    Li, Jian
    Zhu, Anna
    [J]. PROCEEDINGS OF 2020 3RD INTERNATIONAL CONFERENCE ON ROBOT SYSTEMS AND APPLICATIONS, ICRSA2020, 2020, : 40 - 44
  • [6] Imagination Based Sample Construction for Zero-Shot Learning
    Yang, Gang
    Liu, Jinlu
    Li, Xirong
    [J]. ACM/SIGIR PROCEEDINGS 2018, 2018, : 941 - 944
  • [7] Discriminative Learning of Latent Features for Zero-Shot Recognition
    Li, Yan
    Zhang, Junge
    Zhang, Jianguo
    Huang, Kaiqi
    [J]. 2018 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2018, : 7463 - 7471
  • [8] Zero-shot recognition with latent visual attributes learning
    Xie, Yurui
    He, Xiaohai
    Zhang, Jing
    Luo, Xiaodong
    [J]. MULTIMEDIA TOOLS AND APPLICATIONS, 2020, 79 (37-38) : 27321 - 27335
  • [9] Learning Discriminative Latent Attributes for Zero-Shot Classification
    Jiang, Huajie
    Wang, Ruiping
    Shan, Shiguang
    Yang, Yi
    Chen, Xilin
    [J]. 2017 IEEE INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV), 2017, : 4233 - 4242
  • [10] Marginalized Latent Semantic Encoder for Zero-Shot Learning
    Ding, Zhengming
    Liu, Hongfu
    [J]. 2019 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2019), 2019, : 6184 - 6192