Learning driving styles for autonomous vehicles from demonstration

By Markus Kuderer, Shilpa Gulati, Wolfram Burgard. 2015

460 citations recorded by the source corpus.

Subfields: Automotive Engineering, Control and Systems Engineering, Social Psychology

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Automotive Engineering

In Control and Systems Engineering

In Social Psychology

Citations in the corpus

Deep Reinforcement Learning for Autonomous Driving: A Survey (2020, methodology): “Authors of [116] proposed to learn comfortable driving trajectories optimization using expert demonstration from human drivers using Maximum Entropy Inverse RL. Authors of [117] used DQN as the reﬁnement step in IRL to extract the rewards, in an effort learn human-like lane change behavior.”
Adversarial Imitation via Variational Inverse Reinforcement Learning (2018, background): “…function that inculcates the desired behavior can be challenging for many robotic applications, which include learning social-interaction skills (Qureshi et al., 2018), dexterous manipulation (Finn et al., 2016b), autonomous driving (Kuderer et al., 2015), and robotic surgery (Yip & Das, 2017).”
Human-Like Autonomous Car-Following Model with Deep Reinforcement Learning (2018, unknown): “' ( , ) [ max ( ', ')] a Q s a E r Q s a  = + (4)”
Planning for Autonomous Cars that Leverage Effects on Human Actions (2016, background): “This builds on work in social navigation which accounts for interaction potentials with human trajectories [11, 23]: the human and the robot trajectories are jointly planned as the optimum of some reward function in order for everyone to reach their goals and avoid each other.”
Inverse reinforcement learning for intelligent mechanical ventilation and sedative dosing in intensive care units (2019, methodology): “These methods have achieved substantial success in a wide range of applications, from imitation of autonomous driving behaviors [11, 12], control of robot navigation [13] to high dimensional motion analysis [14].”
A Survey of State-Action Representations for Autonomous Driving (2018, background): “This is the case of Kuderer et al. (2015) who use this framework to infer comfortable driving styles from human demonstrations.”
A Survey of End-to-End Driving: Architectures and Training Methods (2020, background): “Comfort includes motion comfort (sickness), apparent safety and level of controllability [86], [88].”
A Survey of Inverse Reinforcement Learning: Challenges, Methods and Progress (2018, background): “Other important examples include boat sailing [13], and learning driving styles [14].”
A Survey of Imitation Learning: Algorithms, Recent Developments, and Challenges (2023, unknown): “Additionally, this process can be inherently unsafe, especially in high-risk applications, such as autonomous driving or aircraft control [49].”
Watch this: Scalable cost-function learning for path planning in urban environments (2016, unknown): “ecent work on IRL lead to signiﬁcant success for long-term planning tasks. Results include improving driving and robotic navigation with focus on interaction of mobile autonomous platforms and humans [17], [18]. Jain et al. [13] addressed learning from direct user interaction focusing on scaling to ”
A Survey of Deep Learning Applications to Autonomous Vehicle Control (2019, background): “The downside of this approach is the time intensive hand-tuning of parameters [20] and the difficulty of”
Recent Advances in Robot Learning from Demonstration (2020, background): “Nevertheless, kinesthetic teaching can be applied in large vehicles, such as cars, where a human takes the driver’s seat and demonstrates a desired behavior by driving (173).”
Federated Learning With Blockchain for Autonomous Vehicles: Analysis and Design Challenges (2020, background): “Most of the existing work ([7]–[9], [12], [14]) that explains and quantifies the performance of blockchain (resp.”
Research Advances and Challenges of Autonomous and Connected Ground Vehicles (2019, background): “With Gaussian distribution, inverse reinforcement learning is widely utilized to capture relevant features about the interaction between dynamic agents and predict social environments [369]– [372].”
Exploring applications of deep reinforcement learning for real-world autonomous driving systems (2019, methodology): “IRL for driving styles : Individual perception of comfort from demonstration is proposed in (Kuderer et al., 2015), where individual driving styles are modeled in terms of a cost function and use feature based inverse reinforcement learning to compute trajectories in vehicle autonomous mode.”
Distributed Multiagent Coordinated Learning for Autonomous Driving in Highways Based on Dynamic Coordination Graphs (2020, methodology): “[37] used feature-based inverse RL to learn the driving policy for a vehicle from observed human style.”