Cost-Effective Research Platform for Child-Robot Interaction Studies Using a Smartphone-Based Humanoid Robot with Double Gesture Arms

Background This research aims to develop a smartphone-based interface robot with dual gesture arms for Child-Robot Interaction (CRI). The robot features two gesture arms powered by shape memory alloy (SMA) actuators, facilitating offering a CRI research platform. Using a smartphone as the system base, functionality, affordability, and internet connectivity were achieved. In the CRI research, these robots serve as tutors for children, maintaining children's motivation, enhancing learning outcomes. The proposed robot offers child-friendly interfaces, incorporates various sensors, and ensures constant internet connectivity. Essential functions include dual gesture arms for natural communication, a locomotion system for mobility, and an expression presentation function for conveying emotions effectively. Research Objectives A CRI research platform equipped with a smartphone-based interface robot is proposed. The achievable functions are defined as the following three features and set them as the research goals of this study. 1. Dual Gesture Arms 2. Mobility Mechanism 3. Expression Presentation Function Methodology (System Design) Since our robot uses the smartphone as a core of the system, it achieves required features with affordable cost. The smartphone-based robot provides a high processing power and incorporates various sensors (including camera, accelerometers, gyroscopes, magnetometers, touch panel, microphone, infrared sensor, atmospheric pressure, GPS, and fingerprint authentication). Additionally, it offers constant internet connectivity and tools for safeguarding users' confidential personal information. Results and Discussion (Experimental Assessment) The integration of each function was carried out to prototype and evaluate the interface robot for the CRI research platform. Each joint of the dual gesture arms achieved movements within 2 seconds, necessary for natural communication with children. The prototype gesture arms driven by SMA has 3 degrees of freedom. Although the yaw axis of the shoulder is omitted, by combining the pitch and roll axes of the shoulder, movements similar to the yaw axis can be achieved. Additionally, a thicker frame structure was designed, resulting in higher visibility of the gesture arms. Conclusion To summarize, our proposal advances the idea of a cost-effective CRI research platform centered around a child learning partner robot, which has dual gesture arms and utilize a smartphone as a core of the system. We plan to deploy this proposed system for actual CRI research and for learning effectiveness measurements.