武汉工程职业技术学院学报

为解决传统农业监测与控制系统依赖云端决策时存在的能耗、延迟与可持续性问题，设计了一种基于ESP32+TinyML的边缘智能自适应环境监测与调控系统。系统在节点侧部署轻量化TinyML模型，用于环境趋势预测与异常识别，有效降低冗余能耗与通信负载；能源方面采用“太阳能+超级电容”方案，结合深度休眠与按需唤醒策略以降低能耗。经过试验，系统实现了节点环境信息自适应监测、趋势预测以及自主调控，为低功耗、预测性与可持续的智慧农业物联网系统提供了一种可行路径。

The decision-making of traditional agricultural monitoring and control systems relies on cloud computing,which has significant shortcomings in terms of energy consumption,latency,and sustainability.To address the above issues,this study designs an edge-intelligent adaptive environment monitoring and control system based on ESP32+TinyML(Tiny Machine Learning).The system deploys a lightweight TinyML model on the node side for environmental trend prediction and anomaly recognition,which can effectively reduce redundant energy consumption and communication load.The system adopts a“solar energy and supercapacitor” solution for energy optimization design,combining deep sleep and ondemand wake-up.After experimentation,the system has achieved adaptive monitoring,trend prediction of node environmental information,and autonomous regulation.This study provides a feasible path for a low-power,predictive,and sustainable smart agricultural IoT system.