智慧家教機器人 ECRobot

GitHub

English

The ECRobot (Empathy Cognition Robot) project investigates the effect of empathetic design on intelligent tutoring systems (ITS). The system assists students in completing linguistic puzzles with LLM-generated, empathetic real-time feedback.

In addition, we developed simulation students based on cognitive architecture. We first examine the fidelity and validity of the simulated students on the error pattern and

Developed individually over 3 months, the system consists of three main components:

1. A customized iPad app (Swift) as the answering interface for linguistic puzzles.
2. An interactive voice-user interface (VUI) allowing participants to discuss tasks with the NAO robot in real time.
3. A FastAPI backend handling database management and connecting all components.

The empathetic feedback algorithm is grounded in Prospect Theory.

We examined the effectiveness and utility of the intelligent tutoring system ECRobot, a realtime multimodal system. For system robustness, we implemented stress tests towards multimodal inputs. On top of that, we utilized simulated students for examining the effectiveness of feedback from the ITS.

中文

ECRobot（情感認知機器人）計畫旨在探討「同理心設計」對智慧教學系統（ITS）的影響，透過大型語言模型（LLM）生成具同理心的即時回饋，協助學生完成語言謎題。

本系統由本人獨立開發，歷時約三個月，包含三個主要模組：

1. iPad 應用程式（Swift 開發）：作為語言謎題的作答介面。
2. 語音使用者介面（VUI）：讓參與者能與 NAO 機器人進行即時對話。
3. FastAPI 後端：負責資料庫管理並整合各模組。

同理心回饋演算法的設計基於展望理論（Prospect Theory）。

除此之外，用於檢驗系統效能的研究已經發表在企業資訊系統（Enterprise Information System）(Yueh et al., 2026)。