Beyond the Session: Centering Teleoperators in Socially Assistive Robot-Child Interactions Reveals the Bigger Picture

Saad Elbeleidy
ACM CSCW
Published in
3 min readOct 2, 2023

This post is a summary of the paper “Beyond the Session: Centering Teleoperators in Socially Assistive Robot-Child Interactions Reveals the Bigger Picture” by Saad Elbeleidy, Terran Mott, Dan Liu, Ellen Yi-Luen Do, Elizabeth Reddy, and Tom Williams.

A flow chart “Deliver Therapy”. The flow chart contains an outer cycle with a label for “weeks or months” with the following connections: Examine then Evaluate then Prepare then Conduct Session then back to Examine. Examine, Evaluate, and Prepare involve Multiple Stakeholders. “Conduct Session” is labeled as “Therapist plus Child” and contains another cycle within it: Examine then Evaluate then Prepare then Deliver Intervention then Examine again. “Conduct Session” is labeled” minutes or hours”
The dual-cycle process of therapy.

In many research labs worldwide, researchers have demonstrated that robots are an effective tool to support therapists and educators working with children. Several organizations are starting to use these robots in therapy with children in practice. While the impact on children is relatively extensively researched, little is known about the people behind the robots; the caregivers controlling these robots.

Robots that assist people through social interaction are called Socially Assistive Robots.

We interviewed 9 people with experience using robots in practice to discuss their needs in providing their services with and without a robot. Here, we discuss what we learned.

Therapy typically follows a dual-cycle process (shown in the flow chart above). When people talk about therapy, they typically think of everything happening within a therapy session itself. We call this the inner cycle. Within the inner cycle of therapy, therapists examine their client, evaluate the client’s needs, prepare an appropriate intervention, and then deliver that intervention. We also learned about the outer cycle, which follows a similar pattern but happens outside of therapy sessions over many days or weeks. In the outer cycle, therapists collaborate with others to perform the same tasks around examining, evaluating, and preparing.

Based on what our participants told us about their needs when using robots, we identified six high level themes:

  • Preparation: therapists need to create or organize the resources they need for therapy.
  • Variety: therapists rely on a variety of content to deliver effective therapy.
  • Awareness: therapists maintain a high level of awareness in therapy to respond to their client and surroundings.
  • Adaptability: therapists adapt within each session to clients’ needs in the moment.
  • Documentation: therapists rely on thorough documentation.
  • Evaluation: therapists continuously evaluate their clients to determine progress and upcoming interventions.
Needs & Difficulties. 1) Preparation a) Therapists overprepare b) Preparation is demanding c) Therapists collaborate. 2) Variety a) Customization b) Relationship building c) Robots provide variety. 3) Awareness a) Therapy demands circumstantial awareness b) Depends on emotional safety c) Robots require attention. 4) Adaptability a) Therapy unpredictable b) Therapists adapt cleverly. 5) Documentation a) Collaborative b) Therapists review documentation. 6) Evaluation a) Institutional b) Iterative

In our paper, we break down each of these categories into several subcategories based on direct quotes from our interviews (seen in the figure above). We also describe several patterns that show how these categories relate to each other and where they fall within the inner and outer cycles. For example, we describe how therapists manage uncertainty to ensure emotional safety:

“Therapists manage uncertainty to ensure emotional safety”. An outer circle labeled “Outer Cycle” and an inner circle labeled “Inner Cycle”. Diagram shows “Awareness — therapy depends on emotional safety” (inner cycle) connected to “Variety — variety is for relationship building” (inner cycle), connected to “Preparation — therapists overprepare (outer cycle) connected to “Adaptability — therapy is unpredictable” (inner cycle), connected to “Awareness therapy demands circumstantial awareness”

In our paper, we present 17 recommendations for developers of robot control interfaces that aim to meet the needs of the people using robots in practice. These can be summarized through three high-level recommendations:

  1. meet therapists’ needs in the moment, e.g., support content that accounts for client-specific preferences.
  2. account for the inner and outer cycle and their relationships, e.g., incorporate collaborative documentation that identifies the documentation author.
  3. move actions to the outer cycle when possible, e.g., separate authoring and operating modes in robot-assisted therapy interfaces.

Above all, we recommend that developers of robots think of their core audience as including robot teleoperators and not solely focus on the end-user of the robot itself.

--

--

Saad Elbeleidy
ACM CSCW

Robot Teleoperation Interface Researcher interested in Machine Learning, Data Visualization, Algorithmic Bias, and Food