Bi3: A Biplatform, Bicultural, Biperson Dataset for Social Robot Navigation

Andrew Stratton1, Phani Teja Singamaneni2, 3, Pranav Goyal1, Rachid Alami2, Christoforos Mavrogiannis1,
1University of Michigan, 2LAAS-CNRS, 3INRIA
Paper arXiv Video Code Data

Abstract

We contribute Bi3, a dataset of social robot navigation among groups of people in a constrained lab space. Compared to prior data collection efforts for social robot navigation, our dataset is unique in that it features: an original experiment design giving rise to close navigation encounters between two humans and a robot; five different navigation algorithms; two different robot platforms; a diverse participant pool of 74 people recruited from two sites in the USA and France; multimodal data streams including 10.5 hours of human and robot ground-truth motion tracks, RGB video, and user impressions over robot performance. Our analysis of the collected crowd behavior through metrics like interaction density and human velocity suggests that Bi3 represents a benchmark of unique diversity and modeling complexity. Thus, Bi3 contributes towards understanding how humans and robots can productively mesh their activities in constrained environments, and can be a resource for training models of human motion prediction and robot control policies for navigation in densely crowded spaces.

Video

Experimental Setup

Our data is collected in a laboratory setting with two humans and one robot. Each navigates a shared workspace completing a block stacking task designed to emulate close-proximity settings such as homes, hospitals, and warehouses.

Prediction Experiment Dataset

Our dataset has diversity in robot platforms and cultural background, being collected at two sites: the University of Michigan in the USA and LAAS-CNRS in France. The full dataset includes 74 participants, totaling 10.5 hours of data.

University of Michigan experimental setup

UM

LAAS-CNRS experimental setup

LAAS

Analysis

We compare our dataset to several existing datasets and find it has the highest standard deviation in human velocity, lowest minimum distance at both sites, and highest average velocity at LAAS. These statistics show the human motion in Bi3 is diverse and challenging.

hmd deanon
hvel deanon

Our dataset contians further diversity in terms of robot motion, which we generated using a model predictive controller leveraging five distinct human motion prediction models. We quantify aspects of the robot motion, and find that the motion differed substantially across sites. Although the motion statistics for CV, CoHAN, and HST are similar, our user impression data for them are very different (see in the paper), validating the importance of our collected user impressions.

rmd
raccel
rvel
rpe

If you find our dataset relevant, please cite it using the reference below:

BibTeX

@inproceedings{stratton2026bi3dataset,
  author    = {Stratton, Andrew and Singamaneni, Phani Teja and Goyal, Pranav and Alami, Rachid and Mavrogiannis, Christoforos},
  title     = {Bi3: A Biplatform, Bicultural, Biperson Dataset for Social Robot Navigation},
  booktitle   = {Proceedings of the IEEE International Conference on Robotics and Automation (ICRA)},
  year      = {2026},
}

User Study

We have also published a separate, detailed work on the human subject study this dataset was created in service of. Check it out here.