Latent Imagination Facilitates Zero-Shot Transfer in Autonomous Racing

被引:18
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
来源
2022 IEEE INTERNATIONAL CONFERENCE ON ROBOTICS AND AUTOMATION, ICRA 2022 | 2022年
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 61 条
[1]  
Abdolmaleki Abbas, 2018, INT C LEARNING REPRE
[2]   Optuna: A Next-generation Hyperparameter Optimization Framework [J].
Akiba, Takuya ;
Sano, Shotaro ;
Yanase, Toshihiko ;
Ohta, Takeru ;
Koyama, Masanori .
KDD'19: PROCEEDINGS OF THE 25TH ACM SIGKDD INTERNATIONAL CONFERENCCE ON KNOWLEDGE DISCOVERY AND DATA MINING, 2019, :2623-2631
[3]   Learning Robust Control Policies for End-to-End Autonomous Driving From Data-Driven Simulation [J].
Amini, Alexander ;
Gilitschenski, Igor ;
Phillips, Jacob ;
Moseyko, Julia ;
Banerjee, Rohan ;
Karaman, Sertac ;
Rus, Daniela .
IEEE ROBOTICS AND AUTOMATION LETTERS, 2020, 5 (02) :1143-1150
[4]   Accurate Mapping and Planning for Autonomous Racing [J].
Andresen, Leiv ;
Brandemuehl, Adrian ;
Hoenger, Alex ;
Kuan, Benson ;
Voedisch, Niclas ;
Blum, Hermann ;
Reijgwart, Victor ;
Bernreiter, Lukas ;
Schaupp, Lukas ;
Chung, Jen Jen ;
Buerki, Mathias ;
Oswald, Martin R. ;
Siegwart, Roland ;
Gawel, Abel .
2020 IEEE/RSJ INTERNATIONAL CONFERENCE ON INTELLIGENT ROBOTS AND SYSTEMS (IROS), 2020, :4743-4749
[5]  
Babu VS, 2020, IEEE INT CON AUTO SC, P1614, DOI [10.1109/case48305.2020.9216949, 10.1109/CASE48305.2020.9216949]
[6]  
Barth-Maron G., 2018, INT C LEARNING REPRE
[7]   An Online Training Method for Augmenting MPC with Deep Reinforcement Learning [J].
Bellegarda, Guillaume ;
Byl, Katie .
2020 IEEE/RSJ INTERNATIONAL CONFERENCE ON INTELLIGENT ROBOTS AND SYSTEMS (IROS), 2020, :5453-5459
[8]   Race driver model [J].
Braghin, F. ;
Chell, F. ;
Melzi, S. ;
Sabbioni, E. .
COMPUTERS & STRUCTURES, 2008, 86 (13-14) :1503-1516
[9]  
Coumans E., 2016, PyBullet, a python module for physics simulation for games, robotics and machine learning
[10]  
Deisenroth M., 2011, P 28 INT C MACHINE L, P465