Deep Latent Competition: Learning to Race Using Visual Control Policies in Latent Space

Schwarting, Wilko; Seyde, Tim; Gilitschenski, Igor; Liebenwein, Lucas; Sander, Ryan; Karaman, Sertaç; Rus, Daniela

doi:10.48550/arxiv.2102.09812

Cited by 4 publications

(4 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Their simulator merges an analytic model with a data-driven dynamic model to be used in the real world with a hierarchical policy structure. Schwarting et al [29] learn a world model in latent space to imagine self-play that reduces sample generation in a multi-agent DRL training scheme. The work of Kalaria et al [30] proposes curriculum learning to utilize a control barrier function that is gradually removed during training to not comprise the final performance.…”

Section: B Mapless Methodsmentioning

confidence: 99%

Residual Policy Learning for Vehicle Control of Autonomous Racing Cars

Raphael

Hoornaert

Caccamo

2023

2023 IEEE Intelligent Vehicles Symposium (IV)

View full text Add to dashboard Cite

The interactive decision-making in multi-agent autonomous racing offers insights valuable beyond the domain of self-driving cars. Mapless online path planning is particularly of practical appeal but poses a challenge for safely overtaking opponents due to the limited planning horizon. Accordingly, this paper introduces RaceMOP, a novel method for mapless online path planning designed for multi-agent racing of F1TENTH cars. Unlike classical planners that depend on predefined racing lines, RaceMOP operates without a map, relying solely on local observations to overtake other race cars at high speed. Our approach combines an artificial potential field method as a base policy with residual policy learning to introduce long-horizon planning capabilities. We advance the field by introducing a novel approach for policy fusion with the residual policy directly in probability space. Our experiments for twelve simulated racetracks validate that RaceMOP is capable of long-horizon decision-making with robust collision avoidance during overtaking maneuvers. RaceMOP demonstrates superior handling over existing mapless planners while generalizing to unknown racetracks, paving the way for further use of our method in robotics. We make the open-source code for RaceMOP available at http://github.com/raphajaner/racemop.

show abstract

Section: B Mapless Methodsmentioning

confidence: 99%

Residual Policy Learning for Vehicle Control of Autonomous Racing Cars

Raphael

Hoornaert

Caccamo

2023

2023 IEEE Intelligent Vehicles Symposium (IV)

View full text Add to dashboard Cite

show abstract

“…To better aid reinforcement learning for exploration, a dynamic model of the learning environment is a viable idea [29], [30]. A world model is learned to train agents through imagined self-play to achieve agility in action sequences [31], [32], [33], and the latent strategies of opponents can be predicted to guide the agent toward appropriate directions [34]. Dynamic model learning also plays a very important role in curiosity [35], [36].…”

Section: Forward Model Learningmentioning

confidence: 99%

A Transferable Legged Mobile Manipulation Framework Based on Disturbance Predictive Control

Yao¹,

Wan²,

Yang³

et al. 2022

Preprint

View full text Add to dashboard Cite

Due to their ability to adapt to different terrains, quadruped robots have drawn much attention in the research field of robot learning. Legged mobile manipulation, where a quadruped robot is equipped with a robotic arm, can greatly enhance the performance of the robot in diverse manipulation tasks. Several prior works have investigated legged mobile manipulation from the viewpoint of control theory. However, modeling a unified structure for various robotic arms and quadruped robots is a challenging task. In this paper, we propose a unified framework disturbance predictive control where a reinforcement learning scheme with a latent dynamic adapter is embedded into our proposed low-level controller. Our method can adapt well to various types of robotic arms with a few random motion samples and the experimental results demonstrate the effectiveness of our method.

show abstract

“…SAC is also widely used [11], [56]- [58]. [10], [59] first learns a latent representation of the world and learns through self-play.…”

Section: B Learning-based Planningmentioning

confidence: 99%

Winning the 3rd Japan Automotive AI Challenge -- Autonomous Racing with the Autoware.Auto Open Source Software Stack

Zang,

Tumu,

Betz

et al. 2022

Preprint

View full text Add to dashboard Cite

The 3rd Japan Automotive AI Challenge was an international online autonomous racing challenge where 164 teams competed in December 2021. This paper outlines the winning strategy to this competition, and the advantages and challenges of using the Autoware.Auto open source autonomous driving platform for multi-agent racing. Our winning approach includes a lane-switching opponent overtaking strategy, a global raceline optimization, and the integration of various tools from Autoware.Auto including a Model-Predictive Controller. We describe the use of perception, planning and control modules for high-speed racing applications and provide experience-based insights on working with Autoware.Auto. While our approach is a rule-based strategy that is suitable for non-interactive opponents, it provides a good reference and benchmark for learning-enabled approaches.

show abstract

Deep Latent Competition: Learning to Race Using Visual Control Policies in Latent Space

Cited by 4 publications

References 25 publications

Residual Policy Learning for Vehicle Control of Autonomous Racing Cars

Residual Policy Learning for Vehicle Control of Autonomous Racing Cars

A Transferable Legged Mobile Manipulation Framework Based on Disturbance Predictive Control

Winning the 3rd Japan Automotive AI Challenge -- Autonomous Racing with the Autoware.Auto Open Source Software Stack

Contact Info

Product

Resources

About