HandoverSim: A Simulation Framework and Benchmark for Human-to-Robot Object Handovers

TL;DR

HandoverSim is a new simulation framework and benchmark for evaluating human-to-robot object handovers, utilizing motion capture data and standardized protocols to assess performance and correlate with real-world results.

Contribution

The paper introduces HandoverSim, a comprehensive simulation benchmark for human-to-robot handovers, including datasets, environments, and evaluation metrics.

Findings

Performance of baseline methods correlates with real-world evaluations

Standardized protocols enable consistent benchmarking

Open-source code facilitates community adoption

Abstract

We introduce a new simulation benchmark "HandoverSim" for human-to-robot object handovers. To simulate the giver's motion, we leverage a recent motion capture dataset of hand grasping of objects. We create training and evaluation environments for the receiver with standardized protocols and metrics. We analyze the performance of a set of baselines and show a correlation with a real-world evaluation. Code is open sourced at https://handover-sim.github.io.