Peter Stone's Selected Publications

Classified by TopicClassified by Publication TypeSorted by DateSorted by First Author Last NameClassified by Funding Source


Multi-Agent Inverse Reinforcement Learning in Real World Unstructured Pedestrian Crowds

Multi-Agent Inverse Reinforcement Learning in Real World Unstructured Pedestrian Crowds.
Rohan Chandra, Haresh Karnan, Negar Mehr, Peter Stone, and Joydeep Biswas.
In IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), October 2025.

Download

[PDF]695.7kB  

Abstract

Social robot navigation in crowded public spaces such as university campuses,restaurants, grocery stores, and hospitals, is an increasingly important area ofresearch. One of the core strategies for achieving this goal is to understandhumans' intent--underlying psychological factors that govern their motion--bylearning how humans assign rewards to their actions, typically via inversereinforcement learning (IRL). Despite significant progress in IRL, learningreward functions of multiple agents simultaneously in dense unstructuredpedestrian crowds has remained intractable due to the nature of the tightlycoupled social interactions that occur in these scenarios e.g. passing,intersections, swerving, weaving, etc. In this paper, we present a newmulti-agent maximum entropy inverse reinforcement learning algorithm for realworld unstructured pedestrian crowds. Key to our approach is a simple, buteffective, mathematical trick which we name the so-called"tractability-rationality trade-off" trick that achieves tractability at the costof a slight reduction in accuracy. We compare our approach to the classicalsingle-agent MaxEnt IRL as well as state-of-the-art trajectory prediction methodson several datasets including the ETH, UCY, SCAND, JRDB, and a new dataset,called Speedway, collected at a busy intersection on a University campus focusingon dense, complex agent interactions. Our key findings show that, on the denseSpeedway dataset, our approach ranks 1st among top 7 baselines with > 2×improvement over single-agent IRL, and is competitive with state-of-the-art largetransformer-based encoder-decoder models on sparser datasets such as ETH/UCY(ranks 3rd among top 7 baselines).

BibTeX Entry

@InProceedings{mairl_crowds,
  author   = {Rohan Chandra and Haresh Karnan and Negar Mehr and Peter Stone and Joydeep Biswas},
  title    = {Multi-Agent Inverse Reinforcement Learning in Real World Unstructured Pedestrian Crowds},
  booktitle = {IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)},
  year     = {2025},
  month    = {October},
  location = {Hangzhou, China},
  abstract = {Social robot navigation in crowded public spaces such as university campuses,
restaurants, grocery stores, and hospitals, is an increasingly important area of
research. One of the core strategies for achieving this goal is to understand
humans' intent--underlying psychological factors that govern their motion--by
learning how humans assign rewards to their actions, typically via inverse
reinforcement learning (IRL). Despite significant progress in IRL, learning
reward functions of multiple agents simultaneously in dense unstructured
pedestrian crowds has remained intractable due to the nature of the tightly
coupled social interactions that occur in these scenarios e.g. passing,
intersections, swerving, weaving, etc. In this paper, we present a new
multi-agent maximum entropy inverse reinforcement learning algorithm for real
world unstructured pedestrian crowds. Key to our approach is a simple, but
effective, mathematical trick which we name the so-called
"tractability-rationality trade-off" trick that achieves tractability at the cost
of a slight reduction in accuracy. We compare our approach to the classical
single-agent MaxEnt IRL as well as state-of-the-art trajectory prediction methods
on several datasets including the ETH, UCY, SCAND, JRDB, and a new dataset,
called Speedway, collected at a busy intersection on a University campus focusing
on dense, complex agent interactions. Our key findings show that, on the dense
Speedway dataset, our approach ranks 1st among top 7 baselines with > 2×
improvement over single-agent IRL, and is competitive with state-of-the-art large
transformer-based encoder-decoder models on sparser datasets such as ETH/UCY
(ranks 3rd among top 7 baselines).
  },
}

Generated by bib2html.pl (written by Patrick Riley ) on Mon Sep 29, 2025 19:22:37