System Architecture of the Autonomous First Aid Robot
The autonomous first aid robot is built upon the ELECFREAKS Future Innovator Kit, utilizing its integrated sensors and actuators. At the core is the Nezha V2 controller, which serves as an interface between the sensing/actuating components and the processing unit. A BBC micro:bit, embedded within the Nezha V2, executes the robot’s primary control software.
Control logic is primarily developed using the MakeCode graphical programming environment, with optional functionality extensions implemented in Python.
Companion Compute Unit & AI Integration
For patient interaction, the system employs a Companion Compute Unit running a Semantic Kernel–based application. This application utilizes a cloud-hosted Large Language Model (LLM) provisioned via Azure OpenAI Services and is implemented in C# on the .NET 8 framework.
In the prototype phase, the Companion Compute Unit is realized with a conventional PC, interfaced to the micro:bit via a serial connection. Future iterations aim to replace this setup with a compact embedded platform, such as a Raspberry Pi.
Due to the cloud-based nature of the LLM, continuous internet connectivity is required for operation.
Robot–AI Communication Interface
The micro:bit and Companion Compute Unit communicate via a serial interface, exchanging abstract commands and analytical results. This exchange enables the robot to perform context-specific actions and engage in patient interaction based on natural language input.
