[RSS 2024]: Expressive Whole-Body Control for Humanoid Robots

Official implementation of "Exbody2: Advanced Expressive Humanoid Whole-Body Control"

Estella is an open-source humanoid robot dedicated to building a lifelike anatomical replicant with expressive design and human-centered interaction.

No description available

Environment-aware humanoid locomotion via Dual-Advantage PPO

This code is designed for genenrating expressive behavior for humanoid robots

The Project Page of paper "Do You Have Freestyle? Expressive Humanoid Locomotion via Audio Control"

SENTIO — The Conscious Empathy Interface A deep-tech humanoid empathy system by DevSora, combining AI cognition (AETHER) and expressive robotics (SENTIO) into a unified platform that perceives, understands, and responds with human-like emotional intelligence.

RoboJackson is an end-to-end framework that teaches humanoid robots expressive dance movements by learning from short real-world videos. This project combines 3D pose estimation, motion retargeting, and reinforcement learning to enable lifelike, whole-body robotic dancing—starting from YouTube Shorts.

Webpage for "Unsupervised human-to-robot motion retargeting via expressive latent space" (Humanoids 2023)

No description available

Enhancing End-user Cognition in Human-Robot Interaction with Expressive Humanoid Robots

No description available

No description available

ESP32-based autonomous humanoid serving robot featuring dual operating modes, IMU-stabilized locomotion, sensor-based obstacle avoidance, expressive OLED eyes, and dashboard-controlled interaction.

ESP32-based autonomous humanoid serving robot featuring dual operating modes, IMU-stabilized locomotion, sensor-based obstacle avoidance, expressive OLED eyes, and dashboard-controlled interaction

Voice-controlled humanoid robot using Arduino Mega 2560, PCA9685 servo driver, and DF2301Q offline voice module. Executes expressive arm, eye, mouth, and shoulder movements based on voice commands via I2C and UART, with GPIO indicators for each action.

This project develops a Graph Neural Network that maps human facial landmarks to servo motor commands, enabling humanoid robots to mirror expressions in real time. Using MediaPipe FaceMesh, PyTorch Geometric, and the i2Head InMoov platform, it advances natural, expressive human–robot interaction.

NAOgest is a software application that explores how the NAO humanoid robot can use expressive hand gestures to improve conflict resolution in customer support. Developed with Webots, the simulation evaluates the effects of verbal and non-verbal behaviors on user trust, empathy, and satisfaction during stressful interactions.