cs.ROAug 14, 2026

THRIVE: Therapeutic Humanoid Robot In Virtual Environment

Authors: Jin Xu, Yu-Ping Chen, Ayanna Howard

Organizations: Georgia Institute of Technology · The Ohio State University · Georgia State University

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

This paper presents THRIVE (Therapeutic Humanoid Robot In Virtual Environment), an at-home rehabilitation platform that integrates a suite of virtual-reality upper-body rehabilitation games, a real-time camera-based motion-tracking system, and a socially interactive robot therapist. The system is designed for therapy and intervention in children with upper-limb motor impairments, which can be improved through consistent, task-specific practice. THRIVE features a set of newly designed, engaging games that target functional reaching, grasping, and object-manipulation movements through customizable popping, hitting, catching, and grabbing tasks, while the camera-based tracking system captures the child's kinematic performance during play. A robot therapist - deployable either as a physical robotic coach or as a remote-presence virtual agent - delivers adaptive, dynamic feedback to motivate the child and guide their movements toward therapeutic goals. THRIVE decouples the therapeutic games from the robot embodiment, extending the platform to support various embodiments and different robots within one modular system. This robot-agnostic design makes THRIVE affordable, scalable, and readily adaptable for sustained use in the home, offering a practical pathway to more consistent and engaging upper-limb therapy for children with motor function impairments.

Figures & tables

Figure 1: System Architecture
Figure 2: Popping game. A demonstration of using both arms simultaneously to pop green bubbles as they appear dynamically on screen.
Figure 3: Hitting game. A demonstration of striking through a row of three to five bubbles in sequence, targeting consecutive reaching movements.
Figure 4: Catching game. An example trajectory illustrating the bubble’s dynamic movement across the screen along both horizontal and vertical axes.
Figure 5: Grabbing game. A demonstration of touching, grabbing, and transporting a bubble from one side of the screen to a designated box on the other side, simulating a grasp-and-release task.
Figure 6: From early prototype to production version of the THRIVE robot
Figure 7: Remote Presence of THRIVE Robot
Figure 8: THRIVE virtual agent, modeled from the CAD design of the physical robot.
Figure 9: Non-verbal communication cues of the THRIVE robot
Figure 10: Celebratory movements to encourage children to play
Appendix figures & tables4 assets

Supplementary material from the paper’s appendix.

Appendix

[Original] Hello. My name is Thrive and I will be playing Super Pop with you today. I will ask you to complete a series of tasks and I would love it if you could follow my instructions. When you are ready, please raise both of your hands as high as you can.
Hi there! I’m Thrive, and today we’re going to play Super Pop! I’ll guide you through some fun tasks, and I’d love for you to follow my instructions. When you’re ready, raise both hands as high as you can!
Hello! My name is Thrive, and I’m super excited to play Super Pop with you today! I have some fun tasks for you, so when you’re ready, lift both hands up as high as possible!
Hey there! I’m Thrive, your game buddy for today! We’re going to play Super Pop, and I’ll give you some fun challenges. Ready? Raise both hands up high!
Hi! I’m Thrive, and I can’t wait to play Super Pop with you! I’ll be giving you some fun tasks—when you’re ready, reach up high with both hands!
Hello, my friend! I’m Thrive, and today we’ll play Super Pop together! I’ll guide you through some awesome tasks. If you’re ready, stretch both hands up high!
Hey! I’m Thrive, and I’m thrilled to play Super Pop with you today! I have some cool challenges for you—when you’re ready, raise both hands as high as you can!
Appendix
Table 1: GPT-4-generated utterance variations: Introduction command. The first row is the original utterance from the SuperPop-NAO platform; the remaining ten rows are GPT-4-generated variations preserving the same meaning.
[Original] Keep up the good work. Move a little faster.
You’re doing great! Let’s pick up the pace a little!
Awesome job! Try moving just a bit faster!
Keep going! Can you speed up a little?
You’re on fire! Let’s go a tiny bit quicker!
Wow, you’re doing amazing! Try moving a bit faster!
Fantastic effort! Let’s see if you can go a little quicker!
Appendix
Table 2: GPT-4-generated utterance variations: “Move faster” command. The first row is the original utterance from the SuperPop-NAO platform; the remaining rows are GPT-4-generated variations preserving the same meaning.
[Original] Fantastic! Let us move at the exact same speed.
Fantastic! Keep moving at this exact speed!
Great job! Maintain this pace!
Awesome! Stay at this speed, just like that!
You’re doing great! Keep your movement steady!
Perfect! Let’s continue at this pace!
Nice work! Keep going at this speed!
Appendix
Table 3: GPT-4-generated utterance variations: “Keep the same speed” command. The first row is the original utterance from the SuperPop-NAO platform; the remaining rows are GPT-4-generated variations preserving the same meaning.
[Original] Fantastic! Wow! Good game. Let us play another one.
Fantastic! Wow! That was a great game! Let’s play another one!
Awesome job! That was so much fun! Ready for another round?
You did amazing! What a great game! Let’s go again!
Wow! That was fantastic! How about another game?
Great work! You rocked that game! Let’s play again!
Super effort! That was fun! Want to play one more?
Appendix
Table 4: GPT-4-generated utterance variations: Game-end feedback command. The first row is the original utterance from the SuperPop-NAO platform; the remaining rows are GPT-4-generated variations preserving the same meaning.
More in this topic

Assistive Robotics

Assistive robotics develops robots and wearable devices that support mobility, daily activities, rehabilitation, and care.

See all 207 papers

Explore similar work

CardsList

  1. See, Learn, Assist: Safe and Self-Paced Robotic Rehabilitation via Video-Based Learning from Demonstration

    Mar 15, 2026Ali Alabbas, Camillo Murgia, Joanne Regan +1Assistive RoboticsRobotic Control

    Faculty of Mechanical Engineering, Atlantic Technological University, Galway, Ireland. · Department of Health and Nutritional Science, Atlantic Technological University, Sligo, Ireland.

  2. Leveraging VR Robot Games to Facilitate Data Collection for Embodied Intelligence Tasks

    Apr 18, 2026Yihan Zhang, Ziyun Huang, Linqi YeRobotic Data Collection

    School of Computer Engineering and Science, Shanghai University, Shanghai, 200444, China · School of Future Technology, Shanghai University, Shanghai, 200444, China

  3. Lucid-XR: An Extended-Reality Data Engine for Robotic Manipulation

    Apr 30, 2026Yajvan Ravan, Adam Rashid, Alan Yu +8Robotic Data CollectionRobotics Simulation

    MIT CSAIL · 2FortyFive Labs · 3Caltech +3