Inclusive Multi-input Depth Interaction: Human Movement as a Medium for Robot Control
Posted by Yanjun Li
Author：
Elena Marie Vella, Kim Vincs, Casey Richardson, John McCormick
Research Area:
Human-Robot Interaction · Depth Camera · Gesture Control · Multi-Human Interaction
Summary:
A depth-camera-based control system that enables one or multiple people to control a robot through natural body movements, without user training or calibration.
Traditional robot control systems often rely on specific input devices or single-user interaction, limiting accessibility and collaborative control. This research explores a more natural and inclusive approach using human body movements.
Key Challenges:
●Supporting multiple body parts as control inputs.
●Enabling simultaneous multi-user interaction.
●Accurately detecting and interpreting body movements.
●Ensuring smooth and responsive robot control.
Solution:
The researchers integrated a Microsoft Azure Kinect depth camera with Unity and ROS to control the Elephant Robotics myAGV. The system converts human body movements into anchor points and uses a multi-input control algorithm to translate these movements into robot motion. It supports both individual and collaborative control without prior user calibration.
