Command-Path Runtime Evaluation of a Multi-Source STM32 Smart-Lighting IoT Node

Abstract

Smart-lighting IoT nodes increasingly receive commands from heterogeneous sources such as edge processes, mobile applications, cloud platforms, and natural-language interfaces. However, an application- or cloud-side command record does not necessarily guarantee embedded-node reception, actuator update, or state synchronization. This paper presents the design and runtime evaluation of a multi-source STM32-based smart-lighting IoT node for command-path testing. The prototype integrates an STM32F407 FreeRTOS execution node, an ESP8266 OneNET bridge, optical and environmental sensors, PWM lighting outputs, and WS2812 ambient-light actuators. Three command sources are implemented: a Python edge process for context-aware lighting strategies, a WeChat MiniApp for remote manual control, and an LLM-assisted ESP32-S3 voice module for natural-language command entry. The evaluation collects runtime evidence for property reporting, command forwarding, STM32-side actuation, state synchronization, and response latency. Prototype tests verify sensor reporting, MiniApp control, LLM-assisted voice control, Python edge strategy execution, and cloud-side state synchronization. The latency samples show sub-second response for MiniApp commands with mean latencies of 0.612–0.655 s, a mean latency of 2.330 s for the LLM-assisted voice path, and a mean latency of 0.236 s for the Python edge strategy path.