Human-centred robotics · Research programme
A multi-paper research programme combining robot kinematics, markerless upper-limb tracking and gesture retargeting for rehabilitation-oriented human–robot interaction.
At a glance
Context
Upper-limb rehabilitation research
My focus
Kinematics, perception and gesture retargeting
Tools
ROS, MATLAB, Gazebo, DeepStream and multi-view vision
The challenge
Rehabilitation exercises depend on repeatable movement, but translating a human arm gesture to a mobile manipulator is not a simple copy operation. Depth ambiguity, occlusion, different limb proportions and TIAGo’s redundant kinematics all affect the motion that the robot can reproduce.
My contribution
I worked on TIAGo kinematic modelling, software implementation, markerless perception, motion retargeting, validation and simulation, in collaboration with Università Politecnica delle Marche and CNR-ISTC.
From perception to robot motion
- Capture upper-limb motion from complementary RGB viewpoints.
- Detect anatomical keypoints and reconstruct their three-dimensional position by triangulation.
- Scale the human geometry to the robot and estimate an initial arm configuration.
- Refine the joint solution using inverse kinematics, then smooth and send the motion through ROS.
The research also studied posture optimisation and static force capability, showing why the configuration of a redundant robot matters as much as the target pose itself.
Evidence and limits
Published studies cover formal kinematic models, markerless tracking, multi-view reconstruction and a real-time mimicry prototype. The gesture-mimicry work was primarily validated in MATLAB and Gazebo; it does not establish clinical efficacy, patient validation or medical-device status.
Selected publications
- Posture Optimization of the TIAGo Highly-Redundant Robot for Grasping Operation — Robotics, 2024.
- AI-Enabled Framework for Augmenting Upper Limb Rehabilitation With a Social Robot — IEEE/ASME MESA, 2024.
- Real-Time Mimicry of Human Arm Gestures on a Single-Arm TIAGo Robot — ASME IDETC/CIE, 2025.
- Markerless Upper Limb Motion Tracking — MetroXRAINE, 2025.
This page presents a research programme, not a clinically validated product.