Project Overview

Intro.

Our project aimed to create child-friendly interaction models to enhance engagement between children and robots. We assembled an interdisciplinary elite team, bringing together researchers from internationally renowned institutions such as MIT and the Institute of Psychology, Chinese Academy of Sciences, along with seasoned experts and creative talents from leading entertainment companies like Disney.

The concept version was launched in April 2021, receiving positive user feedback. Over the following two years, we continuously iterated and optimized the product through development. The project was successfully handed over in June 2023.

This project demonstrated a complete product lifecycle from concept to implementation, achieving innovative breakthroughs in children's AI interaction and generating significant social impact.
Working with my wonderful colleagues, from Harvard, MIT, Disney, Cambridge and Star Wars games.
I was deeply involved in the product launch process, and on Concept Product Launch Day, we proudly posed with the robot

Project Honor & My Impact

The Project was honored with Huawei's Human-Computer Interaction Innovation Award (2022) and was selected as a UNICEF AI for Children featured case study (2023).

We developed a comprehensive solution for social robot interaction and  interactive content creation,accumulated over 10,000 test users.

As a core team member, I led the research and development of Human-Robot Interaction (HRI) systems. Working alongside my colleagues, Summer and Jerry, spearheaded the development of a multimodal interaction and decision-making system based on emotion computing. This system, characterized by its precise emotion recognition and natural interactive responses, has been recognized as one of the most flexible robotic products in the HRI field and has successfully secured patent protection.

On the application front, I worked with game design colleagues. Together, we successfully integrated embodied interaction concepts into various interactive scenarios, including music education and human-robot gaming, significantly enhancing the product's educational value and interactive experience. Post-launch user research demonstrated the product's success with a high Net Promoter Score (NPS) of 35.3%.
Children's AI Research Contribution

The project was included in the UNICEF AI for Children case study, and I co-authored the human-robot interaction part of it

Internal R&D Empowerment Contribution

I led the research and development of emotional interaction, successfully applying for a patent for the team

HCI Innovation Contribution

Team won the HUAWEI Human-Computer Interaction Innovation Contribution Award in 2022

High User Recognition

User net promoter score is 35.3%, which is a relatively high level, and is rated by users as the most flexible robot in human-computer interaction at present.

Product Videos

Product Intro.

XiaoYi Spirit is an intelligent companion robot powered by a dual-chip system with a computing power of 17 TOPS.Featuring 9 precision servos and an OLED display system, it can express over 2,000 facial expressions and natural body movements. Through its smart dual cameras and 5-microphone array, it accurately recognizes faces, movements, emotions, and precisely locates sound sources.Multiple sensor systems enable XiaoYi Spirit to perceive environmental changes in real-time, delivering natural and fluid human-robot interaction.

Emotion Recognition & Emotional Companionship

XiaoYi has its own special story and personality, and can show different levels of emotions in its expressions, making its interactions become more engaging and relatable. By recognizing children's emotions through speech, tone, and facial expressions, XiaoYi responds appropriately and guides children in understanding their feelings, providing timely and positive companionship.

Embodied Multiplayer Interactive Games

We've developed a comprehensive suite of gamified interactive content, thoughtfully designed to cover everything from picture books and music appreciation to language learning and cognitive games. Acting as an intelligent guide, the robot encourages children to explore knowledge through observation, critical thinking, and inquiry, while developing problem-solving skills in an engaging, shared play environment.

Immersive Large-Screen English Learning System

The video demonstrates our robot's innovative English learning feature through seamless TV connectivity and interactive gameplay.

Smart TV Integration:
The robot synchronizes with TV displays to create an immersive large-screen learning environment.

Interactive Gaming Experience:
 - Voice-guided gameplay instructions
 - Natural conversation interactions
 - Touch-based interactive games
 - Real-time feedback and encouragement•

Professional Curriculum Design:
 - Developed by certified language education experts
 - Crafted by experienced game designers
 - Age-appropriate content and difficulty progression• Daily English Conversation:
 - Natural language processing for daily communications
 - Contextual learning scenarios
 - Adaptive conversation patterns

This comprehensive solution transforms traditional English learning into an engaging, gamified experience, creating an immersive environment that naturally motivates children to practice and improve their language skills.

Children's Speech Recognition with VAD Calibration

Leveraging both voice timbre and facial recognition capabilities, the robot excels in managing multi-person conversations. We've also fine-tuned our speech recognition system specifically for children's speech patterns and linguistic habits, ensuring smoother and more intelligent interactions between the robot and childrens.

Sound Source Localization & Face Tracking

By combining sound source detection and facial recognition, the robot can precisely locate users in its surroundings, enabling more responsive and dynamic interactions. Thanks to its highly flexible servo motors, the robot can easily keep up with children's rapid movements and activities.

My Jobs

As a core figure in human-computer interaction design, I lead an innovative project that spans Human-Robot Interaction (HRI) and child cognitive psychology. The most exciting aspect of this project is its unprecedented deep integration of humanities and technology. Here, traditional single-domain knowledge is insufficient to support product design; we must rely on the collective wisdom of interdisciplinary teams to create truly valuable solutions.

Innovation and Breakthroughs

Our project has achieved innovative breakthroughs in multiple areas:

  • Developed a unique emotional interaction model
  • Created a natural and fluid multimodal interaction system
  • Established an intelligent social decision-making framework
  • Designed a rich and diverse interactive content library

User Feedback and Reflections

What moves me most is seeing users establish genuine emotional connections with our robots. In a social media survey, I was delighted to discover bloggers spontaneously interacting daily with our robot "Xiao Yi," gaining a large following on video platforms. These authentic user feedbacks not only make me feel gratified but also deeply aware of the importance and responsibility of our work.

Core Work Content

Part 1

Preliminary Research & Product Strategy

Preliminary research begins with understanding the initial product objectives, followed by in-depth user research to validate the feasibility of value propositions and define target user profiles and usage contexts (including When and Where). The insights gained from user research serve as key foundations for product design, ultimately guiding the development and refinement of product strategy.

My Key Deliverables in Research & Product Strategy

Preliminary Research

Preliminary research begins with understanding the initial product objectives, followed by in-depth user research to validate the feasibility of value propositions and define target user profiles and usage contexts (including When and Where). The insights gained from user research serve as key foundations for product design, ultimately guiding the development and refinement of product strategy.

Beginning

At the start, we identified the key questions that needed to be addressed and made assumptions based on communications with upstream and downstream stakeholders.

Research

Centered around these key questions, we designed two user research themes for exploration.

Valid

Based on the research findings, we revised and supplemented our initial assumptions.

Key Challenge

Based on the relevant conclusions from user research, we have identified key challenges:

  1. How to balance the expectations of children and parents
  2. How to attract children to use it long-term

Robot Product Strategy Development

Based on the relevant conclusions from user research, we have identified key challenges and developed a preliminary product strategy to address these challenges. This includes product goals and value propositions, product solutions, and initial market strategies. This will serve as the core guidance for our subsequent design work.

Key Challenges

Based on the relevant conclusions from user research, we have identified key challenges:

  1. How to balance the expectations of children and parents
  2. How to attract children to use it long-term
Part 2

Design Phase

In the product design phase, we adhere to the core principle of
"Experience Shapes Form, Form Determines Architecture."
The experience dimensions primarily include:
  1. Natural Interaction Experience
  2. Humanized Experience
  3. Gamified Experience
  4. Personalized Experience
These experience requirements cascade through the design process, forming a clear decision-making chain:
  • First, experience requirements shape the product form, including appearance design and interactive behaviors
  • Second, product form guides hardware configuration, such as the selection and specifications of sensors, servos, audio systems, batteries, and chips
  • Finally, these elements collectively define the software architecture and technical breakthrough points
Key Processes in the Design Phase
Take natural interaction as an example:

To address the technical challenges in children's speech recognition (such as non-standard pronunciation, irregular grammar, and environmental noise), we innovatively developed lip-reading technology to effectively tackle real-world application scenarios.

This "Experience First" design philosophy has driven us to:
  • Deeply understand and solve real-world scenarios
  • Accumulate extensive critical technical capabilities
  • Develop systematic product development methodologies
This approach has enabled us to create more intuitive and engaging robotic solutions while building a solid foundation of core technologies and methodologies.

My Key Deliverables in Design Phase

Anthropomorphic Experience Design

In anthropomorphic design, I not only need to balance technical feasibility but also consider how to establish genuine emotional connections between robots and humans. This makes me feel like a director creating life, needing to imbue products with unique personalities and souls. Through carefully designed interaction modes, we enable robots to exhibit warmth, understanding, and care, making them trustworthy companions in users' lives.

Multimodal Natural Interaction Design

Designing multimodal interaction systems is like orchestrating an intricate play, requiring perfect coordination of various interaction modes such as voice, visual, and tactile. In this process, I organize a cross-disciplinary team comprising AI experts, roboticists, psychologists, HRI designers, and artists (including scriptwriters, illustrators, animators, etc.). Each expert brings a unique perspective and expertise, collectively crafting smooth and natural human-machine interactions.

Music Enlightenment Interactive Design

In content design, we are committed to creating interactive experiences that bring lasting joy to children and families. Each interaction scenario is thoughtfully conceived, considering both educational value and entertainment, making the robot an indispensable part of family life.

AI Decision-Making System (Collaboration)

In developing the AI decision-making architecture, I work closely with the technical team, focusing on the robot's social decision-making capabilities. Our biggest challenge is to make robots interact with people "just right" - responding appropriately to the right person at the right time. This requires us to build a complex decision system that comprehensively considers:

  • Real-time assessment of user interaction willingness
  • Precise user profile analysis
  • Construction of the robot's world model
  • Instant observation of environmental information
  • Dynamic prioritization of interaction content
Theoretical Studies
  1. HRI交互理论
  2. 拟人化理论
  3. 儿童认知心理学
Part 3

Release Phase

Through the design and execution of the following release process, we ensured optimal feature presentation in a short timeframe while establishing a solid foundation for subsequent product iterations.

My Key Deliverables in Release Phase

Release and Experience Process Planning

Through the design and execution of the following release process, we ensured optimal feature presentation in a short timeframe while establishing a solid foundation for subsequent product iterations.

Release Strategy Development
1. Develop product release strategy and robot demonstration plan.
2. Optimize feature presentation effectiveness within brief launch events.

Exhibition Hall Experience Design
1. Design robot interactions that autonomously attract attention and provide seamless guidance, considering     exhibition hall noise levels and traffic flow.
2. Engage customer interest and guide them through core functionality experiences.

User Feedback Collection and Analysis

1. Design multiple channels for collecting user feedback during the exhibition period, including on-site     questionnaires and online surveys.
2. Analyze feedback and public opinion to identify product highlights and areas for improvement, providing basis for     future iterative optimization.

Part 4

Extended Research Projects

Based on the technical expertise and design experience accumulated during robot development, I assisted other departments with several specialized research initiatives.

My Key Deliverables in Extended Research Projects

Huawei System-Level Voice Assistant Optimization Strategy

1. How to resolve the contradiction between the serious image of system-level voice assistant and the needs of users    across all age groups for smart speakers?
2. Will introducing the "Xiao Yi Sprite" as a child-specific voice assistant conflict with the existing system-level voice     assistant?
3. How to address the user experience issues caused by the current limited AI capabilities conflicting with Xiao Yi's     young female persona?

Child-Car-Robot Experience Design Strategies

1. Research and investigate children's current in-vehicle behaviors and activities
2. Survey parents' expectations and needs for smart vehicle children-oriented functions
3. Analyze competitors' solutions and approaches to in-vehicle children's experience
4. Develop design strategies to enhance children's in-vehicle experience

My Gains

As a core participant in this innovative project, I feel honored to be involved in shaping the future of human-computer interaction. This work is both challenging and creative, allowing me to fully utilize my professional expertise while constantly pushing boundaries and exploring new possibilities. In the future, I will continue to dedicate myself to promoting the deep integration of technology and humanities, creating more meaningful human-computer interaction experiences.

Future Outlook

As technology advances, robots will play increasingly important roles in people's lives. They will not only be functional assistants but also emotional companions. This brings new challenges and considerations:

  • How to ensure product ethics
  • How to protect user privacy
  • How to assume greater social responsibility

Personal Growth

This project has given me not just professional skill enhancement but also a transformation in thinking:

  • Cultivated interdisciplinary thinking
  • Deepened understanding of the essence of human-computer interaction
  • Improved project management and team collaboration skills