# Learning and Blending Robot Hugging Behaviors in Time and Space

arXiv: https://arxiv.org/abs/2212.01507 · HTML: https://mlsanity.com/paper/2212.01507/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in Robot manipulation imitation learning, one of 13 directions in Offline RL and imitation learning from data.
- Builds on 4 of its 32 references that are in the map.
- First posted on arXiv 3 Dec 2022.

## Placed in

- Offline RL and imitation learning from data / [Robot manipulation imitation learning](https://mlsanity.com/offlinerl/robot-manipulation-imitation-learning/)

## Builds on (most cited first, 4 of 4)

- [Probabilistic Multimodal Modeling for Human-Robot Interaction Tasks](https://mlsanity.com/paper/1908.04955/) (2019)
- [The Six Hug Commandments: Design and Evaluation of a Human-Sized Hugging Robot with Visual and Haptic Perception](https://mlsanity.com/paper/2101.07679/) (2021)
- [Learning Interactive Behaviors for Musculoskeletal Robots Using Bayesian Interaction Primitives](https://mlsanity.com/paper/1908.05552/) (2019)
- [In the Arms of a Robot: Designing Autonomous Hugging Robots with Intra-Hug Gestures](https://mlsanity.com/paper/2202.09935/) (2022)

## Built on by (most cited first, 2 of 2)

- [M4Human: A Large-Scale Multimodal mmWave Radar Benchmark for Human Mesh Reconstruction](https://mlsanity.com/paper/2512.12378/) (2025)
- [Bridging the gap between Learning-to-plan, Motion Primitives and Safe Reinforcement Learning](https://mlsanity.com/paper/2408.14063/) (2024)

## Abstract (arXiv)

We introduce an imitation learning-based physical human-robot interaction algorithm capable of predicting appropriate robot responses in complex interactions involving a superposition of multiple interactions. Our proposed algorithm, Blending Bayesian Interaction Primitives (B-BIP) allows us to achieve responsive interactions in complex hugging scenarios, capable of reciprocating and adapting to a hugs motion and timing. We show that this algorithm is a generalization of prior work, for which the original formulation reduces to the particular case of a single interaction, and evaluate our method through both an extensive user study and empirical experiments. Our algorithm yields significantly better quantitative prediction error and more-favorable participant responses with respect to accuracy, responsiveness, and timing, when compared to existing state-of-the-art methods.