A Multi-Layer Architecture for Intelligent Chatbots in IoT Robots: Design and Implementation

This thesis presents the design and implementation of Rima, an intelligent chatbot system tailored for IoT robots. The system utilizes a three-layer architecture that integrates Language Understanding Intelligent Service (LUIS), a retrieval-based chatbot (AIML), and a generative model (Seq2Seq), enabling both precise command recognition and fluid natural language conversation. Developed through a cross-platform approach, Rima operates across diverse client environments and provides API support for seamless integration into IoT hardware. Furthermore, by enhancing AIML tags, the system incorporates online learning capabilities, allowing the robot to update its common-sense knowledge base in real-time during interactions. Experimental results demonstrate that this multi-layer hybrid structure performs effectively in both task execution and open-domain chat, proving its high practical value for the IoT ecosystem.

Thesis