Building Bridges with Bots: A Robot Companion for Newcomer Kids

by

Author: Denis Avetisyan

Researchers co-designed Maple, a social robot, to help young immigrants learn English and adapt to a new culture through community-based literacy programs.

The co-design process, structured around the principles of Design Thinking, progressed through five phases-empathizing via shadowing and observational analysis, defining challenges and user needs through transcript analysis, ideating robot roles and lesson activities, prototyping an interactive system with iterative refinement, and testing through tutor playtesting, interviews, and child case studies-to forge a human-centered design.
The co-design process, structured around the principles of Design Thinking, progressed through five phases-empathizing via shadowing and observational analysis, defining challenges and user needs through transcript analysis, ideating robot roles and lesson activities, prototyping an interactive system with iterative refinement, and testing through tutor playtesting, interviews, and child case studies-to forge a human-centered design.

This paper details the co-design and development of Maple, a socially assistive robot supporting English language acquisition and cultural orientation for newcomer children within tutor-mediated community literacy settings.

Community literacy programs supporting young newcomers often struggle with limited resources to provide individualized language and cultural support. This paper details the co-design process and development of Maple, a table-top, peer-like Socially Assistive Robot (SAR) explored in ‘You’ve got a friend in me”: Co-Designing a Peer Social Robot for Young Newcomers’ Language and Cultural Learning’, intended to function as a practice partner within tutor-mediated sessions. Through co-design interviews and iterative prototyping, we developed a system incorporating multimodal scaffolding and embedded formative assessments to support both language socialization and attentive engagement. How might such tutor-in-the-loop SARs be optimally integrated into authentic community programs to foster inclusive learning experiences?

The Weight of Words: Integrating Newcomers Through Language

The process of acquiring a new language is rarely straightforward for children, but newcomer children face a uniquely complex set of hurdles beyond simple linguistic acquisition. Simultaneously adapting to a new educational system, navigating unfamiliar social cues, and processing a different cultural landscape creates significant cognitive and emotional demands. This confluence of challenges can impede not only English as a Second Language (ESL) development, but also a child’s overall sense of belonging and ability to fully participate in Canadian society. Successful integration hinges on more than just vocabulary and grammar; it requires a delicate balance of linguistic competence and cultural understanding, and when one area lags, it can negatively impact the other, potentially leading to feelings of isolation and hindering academic progress.

While community literacy programs represent a vital resource for newcomer children, their broad scope often limits the provision of truly individualized language support. These programs, frequently designed to address foundational literacy skills for diverse populations, may struggle to offer the sustained, engaging practice necessary for full language socialization-the process by which children learn to use language appropriately in social contexts. The need extends beyond grammatical correctness; it encompasses understanding nuanced communication, participating in classroom discussions, and building confidence in everyday interactions. Without sufficient opportunities for personalized application and feedback, children may progress through initial literacy stages but still encounter barriers to fully participating in Canadian school and community life, hindering their overall integration and long-term success.

Successful integration for newcomer children extends far beyond simply learning English; it demands a comprehensive understanding of Canadian social customs and full participation in community activities. Linguistic competence is undoubtedly crucial, but navigating unspoken rules, understanding cultural nuances, and building social connections are equally vital for a sense of belonging and well-being. This need is increasingly urgent given Canada’s significant population growth – a surge of 1,271,872 new residents between January 2023 and January 2024, primarily fueled by international migration. These children require support not only in language acquisition, but also in deciphering the social landscape, enabling them to thrive as active and engaged members of Canadian society.

Maple: A Peer-Like Presence for Language Practice

Maple is a table-top humanoid robot developed for use as a practice partner within Community Literacy Programs. The robot’s physical form factor was specifically chosen to facilitate social interaction and encourage a peer-like dynamic with participating children. Unlike fully autonomous tutoring systems, Maple is designed to operate within existing literacy programs, providing structured, repetitive practice opportunities that complement the instruction provided by human tutors. The robot measures approximately 30cm in height and incorporates features such as expressive eyes and articulated arms to enhance engagement and facilitate multimodal interaction. Development prioritized a design that is both robust for repeated use in community settings and approachable for young learners.

Maple operates on a Tutor-Mediated Interaction model, prioritizing the human tutor’s role in literacy instruction. The robot is not intended to replace the tutor but to function as a supplementary tool providing consistent, repetitive practice in foundational skills. This approach allows tutors to focus on more complex aspects of language learning, such as nuanced conversation and creative expression, while Maple handles tasks requiring high-frequency repetition, like phonics drills or vocabulary review. The tutor maintains control of the session, selecting activities and providing individualized support, with Maple responding to cues and delivering pre-programmed exercises as directed.

The development of Maple incorporated a Co-Design Methodology to directly address the requirements of both literacy tutors and newcomer children. This process involved gathering input from six second language (L2) acquisition experts through dedicated group interviews. These interviews informed design decisions related to interaction style, activity structure, and the robot’s overall role as a supplemental practice tool. The resulting design prioritizes tutor control, allowing them to integrate Maple into existing lesson plans, while providing children with repetitive, engaging opportunities to practice language skills in a low-pressure environment. The Co-Design approach ensured the robot’s functionality aligned with established pedagogical practices and the specific needs of the target user groups.

Multimodal scaffolding within the Maple robot’s interactions utilizes both visual and interactive elements to improve learner engagement and facilitate language acquisition. This approach incorporates visual cues, such as images and highlighted text, to support comprehension and provide contextual information. Interactive activities, including question prompts, fill-in-the-blank exercises, and simple games, encourage active participation and provide opportunities for practice. The combination of these modalities aims to reduce cognitive load and provide multiple pathways for understanding, ultimately supporting language development for newcomer children participating in literacy programs.

The Maplerobotic system comprises a robotic prototype integrated with a system architecture designed for [latex] ext{insert specific functionality or application here}[/latex].
The Maplerobotic system comprises a robotic prototype integrated with a system architecture designed for [latex] ext{insert specific functionality or application here}[/latex].

Technical Foundations: Bringing Maple to Life

Maple’s core functionality is implemented using the Robotic Operating System (ROS), a widely adopted open-source framework for building robot software. ROS provides hardware abstraction, low-level device control, message passing, and package management, facilitating modular development and code reuse. This architecture allows for integration of diverse components, including speech synthesis, facial expression rendering, and assessment tools, within a unified system. The use of ROS also enables Maple to operate on a variety of hardware platforms and simplifies future expansion and maintenance through its established community and extensive library of pre-built functionalities.

Speech synthesis in Maple is achieved through integration with Kokoro-TTS, a text-to-speech system designed for high intelligibility and natural prosody. Kokoro-TTS utilizes a neural network architecture trained on a large corpus of human speech data, enabling it to generate audio with minimal robotic artifacts. The system supports multiple voices and allows for dynamic control over speech parameters, including rate, pitch, and volume, facilitating nuanced and contextually appropriate communication during interactions. This ensures that Maple’s verbal responses are not only understandable but also contribute to a more engaging and realistic user experience.

Facial expressions for Maple are generated using PyLips, a Python library designed for real-time lip synchronization and facial animation. The system operates by manipulating Facial Action Units (AUs), which represent specific muscular movements of the face. By combining different AUs – such as raising the inner brow (AU1) or pulling the lip corners back (AU12) – PyLips constructs a wide range of expressions. This AU-based approach allows for nuanced control over facial animation, moving beyond simple pre-defined expressions and enabling Maple to convey subtle emotional cues and enhance user engagement during interactions.

Embedded Assessment within Maple utilizes data collected during interactive, game-like activities to gauge learner performance. This approach moves beyond traditional testing methods by continuously monitoring responses and behaviors – such as response time, accuracy, and interaction patterns – as the learner engages with the activity. The resulting data is then analyzed to provide insights into areas of strength and weakness, identifying specific skills requiring further development. Critically, this assessment is designed to be non-intrusive, occurring seamlessly within the playful experience to avoid interrupting the learning process or inducing test anxiety. The system generates reports detailing progress and informs adaptive learning pathways, allowing Maple to tailor future interactions to individual learner needs.

Maple's facial expression system was updated from a previous implementation [38] to the current design [11] to improve realism and expressiveness.
Maple’s facial expression system was updated from a previous implementation [38] to the current design [11] to improve realism and expressiveness.

Beyond Fluency: Cultivating Connection and Community

At the heart of Maple’s methodology lies a commitment to narrative engagement, recognizing that stories are not merely entertainment, but powerful tools for learning and cultural connection. The program centers around carefully designed activities that immerse children in meaningful narratives, allowing them to practice language skills within a context that feels authentic and relatable. This approach moves beyond rote memorization and traditional drills, instead fostering a deeper understanding of language through active participation and emotional resonance. By becoming invested in the unfolding of a story, children naturally absorb new vocabulary and grammatical structures, while simultaneously gaining insights into Canadian culture and societal norms. This immersive experience cultivates not only linguistic competence, but also a sense of belonging and confidence in navigating real-world interactions.

Children learning a new language benefit significantly when instruction is embedded within compelling narratives, allowing them to move beyond rote memorization and truly understand how language functions in real-world contexts. This approach fosters pragmatic competence – the ability to use language appropriately in various social situations – by exposing learners to authentic dialogues, cultural nuances, and implied meanings. When language practice is linked to stories reflecting Canadian life, traditions, and values, children don’t just learn what to say, but how and when to say it within a Canadian cultural framework, greatly enhancing their ability to connect with the community and participate fully in everyday interactions.

The Maple program prioritizes a nurturing atmosphere where children can confidently practice speaking and interacting in a new language. Recognizing that language acquisition is deeply intertwined with emotional wellbeing, the program deliberately cultivates a safe and supportive environment to mitigate anxiety and build self-assurance. This approach acknowledges that children are more likely to actively participate and retain information when they feel secure and comfortable taking risks. By focusing on positive reinforcement and creating opportunities for successful communication, Maple helps children overcome the hesitation often associated with learning a new language, ultimately fostering not just linguistic competence, but also the emotional resilience necessary for successful integration and social participation.

The benefits of Maple’s approach extend significantly beyond simply acquiring linguistic skills; it actively cultivates language socialization, a process crucial for successful community integration. Recognizing that language is inextricably linked to culture and social practices, the program prepares newcomers not just to speak a language, but to use it appropriately within a Canadian context. This is particularly relevant given recent demographic shifts; in 2023, the OECD recorded 6.5 million new permanent immigrants, and international migration accounted for 98% of Canada’s population growth. By fostering pragmatic competence and confidence in real-world interactions, Maple aims to support these new residents in navigating daily life, building relationships, and fully participating in Canadian society, thereby easing the challenges inherent in adapting to a new culture and environment.

The development of Maple, as detailed in the study, prioritizes a carefully considered simplicity. It’s a deliberate crafting of interaction, mirroring the belief that impactful technology doesn’t require needless complexity. This echoes Ada Lovelace’s insight: “The Analytical Engine has no pretensions whatever to originate anything.” Maple isn’t intended to independently teach language or culture; rather, it functions as a scaffold, a tool to enhance the existing tutor-mediated learning process. The robot’s role is to augment, not replace, human interaction, facilitating a more engaging and effective cultural orientation for newcomer children. The emphasis on multimodal scaffolding, central to the project, directly supports this assistive, rather than generative, function.

Further Refinements

The presented work, while a necessary step, reveals the persistent difficulty in translating the nuance of social interaction into algorithmic form. Maple functions as a scaffold, but scaffolding, by its nature, must eventually be removed. The true measure of success will not be the robot’s capabilities, but the degree to which it fosters independence – a capacity for language acquisition and cultural integration that extends beyond its presence. This necessitates a shift in focus: from perfecting the robot, to perfecting the metrics by which its utility is judged.

Current evaluations often rely on easily quantifiable gains in vocabulary or grammatical correctness. These are shadows of the thing itself. Future research should prioritize the assessment of more elusive qualities – a newcomer’s sense of belonging, their capacity to navigate social ambiguity, their ability to initiate and maintain relationships. These are not problems for robots to solve, but conditions to be observed and, perhaps, gently encouraged.

The emphasis on a tutor-mediated approach, while pragmatic, also implies a limitation. The aspiration should not be to replace human connection, but to augment it. The question remains: can a robot, however thoughtfully designed, genuinely facilitate the complex process of cultural learning, or does it merely offer a technologically advanced form of mimicry? The answer, predictably, lies not in the technology itself, but in the humility with which it is deployed.

Original article: https://arxiv.org/pdf/2603.18804.pdf

Contact the author: https://www.linkedin.com/in/avetisyan/

See also:

2026-03-21 03:09