39,487 Hits in 4.8 sec

End-to-end Driving via Conditional Imitation Learning [article]

Felipe Codevilla, Matthias Müller, Antonio López, Vladlen Koltun, Alexey Dosovitskiy
2018 arXiv   pre-print
We propose to condition imitation learning on high-level command input.  ...  Deep networks trained on demonstrations of human driving have learned to follow roads and avoid obstacles. However, driving policies trained via imitation learning cannot be controlled at test time.  ...  Antonio and Felipe also thank Germán Ros who proposed to investigate the benefits of introducing route commands into the end-to-end driving paradigm during his time at CVC.  ... 
arXiv:1710.02410v2 fatcat:4oq5bjtse5du3bnagklmho2ste

Safer End-to-End Autonomous Driving via Conditional Imitation Learning and Command Augmentation [article]

Renhao Wang, Adam Scibior, Frank Wood
2020 arXiv   pre-print
Imitation learning is a promising approach to end-to-end training of autonomous vehicle controllers.  ...  The main contribution of this work is a recipe for building controllable imitation driving agents that improves upon multiple aspects of the current state of the art relating to robustness and interpretability  ...  The dominant paradigm for end-to-end training of driving agents is imitation learning (IL) from human demonstration [1] , [2] , [3] , [4] , [5] , [6] , [7] .  ... 
arXiv:1909.09721v3 fatcat:etul35wfbzfqlhykn4ezigfjsq

CARLA: An Open Urban Driving Simulator [article]

Alexey Dosovitskiy, German Ros, Felipe Codevilla, Antonio Lopez, Vladlen Koltun
2017 arXiv   pre-print
We use CARLA to study the performance of three approaches to autonomous driving: a classic modular pipeline, an end-to-end model trained via imitation learning, and an end-to-end model trained via reinforcement  ...  learning.  ...  The authors are particularly grateful to Nestor Subiron, the principal programmer, and Francisco Perez, the lead digital artist, for their tireless work.  ... 
arXiv:1711.03938v1 fatcat:ol6s3znpvjc2rmbvv7zvxizixi

A Survey of Deep Reinforcement Learning Algorithms for Motion Planning and Control of Autonomous Vehicles [article]

Fei Ye, Shen Zhang, Pin Wang, Ching-Yao Chan
2021 arXiv   pre-print
Therefore, this paper also presents a growing trend of work that falls into the end-to-end approach, which typically offers better performance and smaller system scales.  ...  In this survey, we systematically summarize the current literature on studies that apply reinforcement learning (RL) to the motion planning and control of autonomous vehicles.  ...  Since a vehicle trained end-to-end to imitate an expert cannot be controlled at test time (i.e., cannot take a specific turn at an upcoming intersection), a condition imitation learning approach was proposed  ... 
arXiv:2105.14218v2 fatcat:27glt4i4lfhg3j4ozjrlsq6i3e

A Survey of End-to-End Driving: Architectures and Training Methods [article]

Ardi Tampuu, Maksym Semikin, Naveed Muhammad, Dmytro Fishman and Tambet Matiisen
2020 arXiv   pre-print
We review the learning methods, input and output modalities, network architectures and evaluation schemes in end-to-end driving literature.  ...  Autonomous driving is of great interest to industry and academia alike. The use of machine learning approaches for autonomous driving has long been studied, but mostly in the context of perception.  ...  ACKNOWLEDGMENTS The authors would like to thank Hannes Liik for fruitful discussions.  ... 
arXiv:2003.06404v1 fatcat:ekb4g7waa5fyldfaxhgnb3a5xm

CIRL: Controllable Imitative Reinforcement Learning for Vision-Based Self-driving [chapter]

Xiaodan Liang, Tairui Wang, Luona Yang, Eric Xing
2018 Lecture Notes in Computer Science  
To our knowledge, this is the first successful case of the learned driving policy by reinforcement learning in the high-fidelity simulator, which performs better than supervised imitation learning.  ...  Autonomous urban driving navigation with complex multi-agent dynamics is under-explored due to the difficulty of learning an optimal driving policy.  ...  A line of researches [2, 35, 15, 4, 24, 13] for learning policies follow the end-to-end imitation learning that directly maps sensor inputs to vehicle control commands via supervised training on large  ... 
doi:10.1007/978-3-030-01234-2_36 fatcat:ogluc4r4jnglljbw4liklwednu

CIRL: Controllable Imitative Reinforcement Learning for Vision-based Self-driving [article]

Xiaodan Liang, Tairui Wang, Luona Yang, Eric Xing
2018 arXiv   pre-print
To our knowledge, this is the first successful case of the learned driving policy through reinforcement learning in the high-fidelity simulator, which performs better-than supervised imitation learning  ...  Autonomous urban driving navigation with complex multi-agent dynamics is under-explored due to the difficulty of learning an optimal driving policy.  ...  A line of researches [8, 9, 10, 11, 12, 13] for learning policies follow the end-to-end imitation learning that directly maps sensor inputs to vehicle control commands via supervised training on large  ... 
arXiv:1807.03776v1 fatcat:2pa3slj255c77ngk76tvwtufmq

RoboBus: A Diverse and Cross-Border Public Transport Dataset

Georgios Varisteas, Raphael Frank, Francois Robinet
2021 2021 IEEE International Conference on Pervasive Computing and Communications Workshops and other Affiliated Events (PerCom Workshops)  
We use an end-to-end autonomous driving approach that relies on imitation learning as use case example for the dataset.  ...  Academic datasets are an important source of information to validate and benchmark novel research concepts.  ...  The authors would also like to thank Voyages Emile Weber Luxembourg for their support and for granting us access to their bus.  ... 
doi:10.1109/percomworkshops51409.2021.9431129 fatcat:ua7bdhriqjhxrbg6lljggcfgdi

Ignition: An End-to-End Supervised Model for Training Simulated Self-Driving Vehicles [article]

Rooz Mahdavian, Richard Diehl Martinez
2018 arXiv   pre-print
We introduce Ignition: an end-to-end neural network architecture for training unconstrained self-driving vehicles in simulated environments.  ...  Importantly, we never explicitly train the model to detect road features like the outline of a track or distance to other cars; instead, we illustrate that these latent features can be automatically encapsulated  ...  Since then, these advancements in imitation learning have been applied to self-driving algorithms.  ... 
arXiv:1806.11349v1 fatcat:y6kcifz3lvaq5d2hn7z3dip6om

Exploring the Limitations of Behavior Cloning for Autonomous Driving

Felipe Codevilla, Eder Santana, Antonio Lopez, Adrien Gaidon
2019 2019 IEEE/CVF International Conference on Computer Vision (ICCV)  
Behavior cloning in particular has been successfully used to learn simple visuomotor policies end-to-end, but scaling to the full spectrum of driving behaviors remains an unsolved problem.  ...  In contrast, imitation learning can, in theory, leverage data from large fleets of human-driven cars.  ...  We use an expanded formulation for self-driving cars called Conditional Imitation Learning, CIL [12] .  ... 
doi:10.1109/iccv.2019.00942 dblp:conf/iccv/CodevillaSLG19 fatcat:sbmedxvpwnhjtk6oajprvnw6mi

Navigating an Automated Driving Vehicle via the Early Fusion of Multi-Modality

Malik Haris, Adam Glowacz
2022 Sensors  
On the other hand, we used the end-to-end driving method to assign raw sensor data directly to vehicle control signals.  ...  The latter is less well-studied but is becoming more popular since it is easier to use. This article focuses on end-to-end autonomous driving, using RGB pictures as the primary sensor input data.  ...  In order to clone human drivers' driving behavior, we can learn a deterministic policy network F via conditional imitation learning.  ... 
doi:10.3390/s22041425 pmid:35214327 pmcid:PMC8878300 fatcat:g7ti567yn5h6jbxbazzeoprnt4

Urban Driving with Conditional Imitation Learning [article]

Jeffrey Hawke, Richard Shen, Corina Gurau, Siddharth Sharma, Daniele Reda, Nikolay Nikolov, Przemyslaw Mazur, Sean Micklethwaite, Nicolas Griffiths, Amar Shah, Alex Kendall
2019 arXiv   pre-print
As our main contribution, we present an end-to-end conditional imitation learning approach, combining both lateral and longitudinal control on a real vehicle for following urban routes with simple traffic  ...  Prior work has studied imitation learning (IL) for autonomous driving with a number of limitations.  ...  [7] learn longitudinal and lateral control via conditional imitation learning, following route commands on a remote control car in a static environment.  ... 
arXiv:1912.00177v2 fatcat:57rjnotxfbak3m433iyjvphkuq

Uncertainty-Aware Data Aggregation for Deep Imitation Learning [article]

Yuchen Cui, David Isele, Scott Niekum, Kikuo Fujimura
2019 arXiv   pre-print
In this work, we present the uncertainty-aware imitation learning (UAIL) algorithm for improving end-to-end control systems via data aggregation.  ...  UAIL applies Monte Carlo Dropout to estimate uncertainty in the control output of end-to-end systems, using states where it is uncertain to selectively acquire new training data.  ...  In this work, we propose an active online imitation learning algorithm for deep end-to-end control systems.  ... 
arXiv:1905.02780v1 fatcat:vmmcu6jxujbq3f5doopej2gcr4

A Probabilistic Framework for Imitating Human Race Driver Behavior

Stefan Lockel, Jan Peters, Peter Van Vliet
2020 IEEE Robotics and Automation Letters  
Experiments in a simulated car racing setting show considerable advantages in imitation accuracy and robustness compared to other imitation learning algorithms.  ...  However, unique driving styles, inconsistent behavior, and complex decision processes render it a challenging task, and existing approaches often lack variability or robustness.  ...  This could be achieved via conditioning of the learned ProMP or by situative modulation of the phase velocityż.  ... 
doi:10.1109/lra.2020.2970620 fatcat:qfbruaoozva6rmyoxf73gtnrb4

MPC-based Imitation Learning for Safe and Human-like Autonomous Driving [article]

Flavia Sofia Acerbo, Jan Swevers, Tinne Tuytelaars, Tong Duy Son
2022 arXiv   pre-print
Imitation learning (IL) algorithms serve this purpose, but struggle to provide safety guarantees on the resulting closed-loop system trajectories.  ...  With this strategy, IL is performed in closed-loop and end-to-end, through parameters in the MPC cost, model or constraints.  ...  Acknowledgements This work was supported by the Flemish Agency for Innovation and Entrepreneurship (VLAIO) in the context of MIMIC (huMan IMItation for autonomous driving Comfort) Baekeland Mandaat [Project  ... 
arXiv:2206.12348v1 fatcat:226h7m4hkff7dptawx4j4x3fxy
« Previous Showing results 1 — 15 out of 39,487 results