Socially compliant mobile robot navigation via inverse reinforcement learning

By Henrik Kretzschmar, Markus Spies, Christoph Sprunk, Wolfram Burgard. 2016

511 citations recorded by the source corpus.

Subfields: Automotive Engineering, Computer Vision and Pattern Recognition, Ocean Engineering

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 Computer Vision and Pattern Recognition

In Ocean Engineering

Citations in the corpus

A Survey of Deep Network Solutions for Learning Control in Robotics: From Reinforcement to Imitation (2016, positive): “…formulation of IRL offers an efﬁcient solution for learning policies for socially compliant navigation (Okal and Arras, 2016; Pfeiffer et al., 2016; Kretzschmar et al., 2016), where the agent needs to not only avoid collisions with static obstacles but also to behave in a socially compliant…”
Socially aware motion planning with deep reinforcement learning (2017, background): “human demonstration (teleoperation) [16], and a probability distribution over the set of joint trajectories with nearby pedestrians [2], [17].”
Deep Local Trajectory Replanning and Control for Robot Navigation (2019, unknown): “Generally, most methods for social navigation explicitly model interactions among agents or social conventions [5, 31, 7, 32, 33, 34, 35, 36, 37, 38, 39].”
Core Challenges of Social Robot Navigation: A Survey (2021, methodology): “Kretzschmar et al. [125] incorporate a featurization that captures discrete ( e.g. , passing from the right or left side) and continuous features of cooperation ( e.g. , speed adjustment) into the general IRL paradigm.”
Planning and Decision-Making for Autonomous Vehicles (2018, background): “Terms for affordance, for progress, and to penalize close distances to other agents can also be included in their joint cost function (88).”
Deep-learning in Mobile Robotics - from Perception to Control Systems: A Survey on Why and Why not (2016, background): “…learning, such as inverted helicopter ﬂight learning (Kim et al., 2003; Ng et al., 2006), gliding aircraft control (Chung et al., 2015), mobile robot navigation (Kretzschmar et al., 2016), modular robot control (Varshavskaya et al., 2008) and biped walking (Benbrahim and Franklin, 1997).”
Theory of mind as inverse reinforcement learning (2019, background): “In robotics, RL planners that integrate IRL can predict where pedestrians are headed and preemptively adjust their plan to avoid collisions (Figure 2a; [20,21]).”
Reinforced Imitation: Sample Efficient Deep Reinforcement Learning for Mapless Navigation by Leveraging Prior Demonstrations (2018, background): “[11] present approaches for navigation in dynamic environments based on IRL.”
Decentralized non-communicating multiagent collision avoidance with deep reinforcement learning (2016, methodology): “cooperative approaches [16]–[18] has been proposed, which first infers the other agents’ intents (e.”
Crowd-Robot Interaction: Crowd-Aware Robot Navigation With Attention-Based Deep Reinforcement Learning (2018, methodology): “In order to navigate through a dense crowd in a socially compliant manner, robots need to understand human behavior and comply with their cooperative rules [8]–[11].”
A Survey of Inverse Reinforcement Learning: Challenges, Methods and Progress (2018, methodology): “er, to avoid an expensive discretization in large continuous spaces, crp-bnirl approximates the demonstration likelihood by comparing actions to an existing closed-loop controller. Kretzschmar et al. [11] apply maxentirl to learn the probability distribution over navigation trajectories of interacti”
Social behavior for autonomous vehicles (2019, unknown): “The optimal control policy of the best-response game with learned rewards yields a human-imitating policy (9, 33, 34).”
Deep-Learned Collision Avoidance Policy for Distributed Multiagent Navigation (2016, background): “The local trajectory should be collision-free with other agents and obstacles, at least within the time horizon τ .”
Virtual-to-real deep reinforcement learning: Continuous control of mobile robots for mapless navigation (2017, methodology): “[15] used inverse reinforcement learning methods to make robots interact with humans in a socially compliant way.”
Watch this: Scalable cost-function learning for path planning in urban environments (2016, background): “IRL specifically addresses the case where, instead of the reward structure, a set of expert demonstrations D = {ς1, ς2, ..., ςN} are supplied.”
A Survey of State-Action Representations for Autonomous Driving (2018, background): “Thus, Kretzschmar et al. (2014, 2016); Pfei(cid:27)er et al. (2016) use cubic splines to represent the future trajectory of each moving entity.”
From perception to decision: A data-driven approach to end-to-end motion planning for autonomous ground robots (2016, methodology): “[11] use the maximum entropy inverse reinforcement learning techniques to interaction-aware motion models.”