AIoT-Powered Autonomous Agricultural Robot
Nguyen Tien Hoang Dung (Class 12A2) and Nguyen Phi Yen (Class 11A4), pupils at Tra On High School in Vinh Long Province, developed an autonomous agricultural robot under the guidance of their teacher, Tran Trong Huu. The robot combines artificial intelligence (AI), Internet of Things (IoT) technologies and computer vision to monitor crops, recognise weeds and plant disease, and remove weeds with a laser instead of chemicals.
Vegetable farming faces several practical problems: reliance on chemical herbicides, labour-intensive weeding, crop diseases that develop quickly, and a lack of accurate digital tools for monitoring field conditions. The two pupils set out to replace manual observation and chemical weed control with a single platform that brings together automation, AI, sensing and renewable energy.
HOW IT WORKS
The robot combines a self-driving platform, image recognition, field sensors and a laser:
- Autonomous mobility and control: A four-wheel platform with tank steering moves forward and backward, turns and rotates in place. A Raspberry Pi 5 serves as the central processor and communicates with two ESP32 controllers for motion, CNC control, servos, limit switches and IoT data collection.
- AI vision and laser weed control: OpenCV analyses image characteristics to identify initial crop and weed areas, and a deep-learning model such as YOLO classifies crops, weeds and plant diseases. A three-axis CNC mechanism positions a laser diode over each detected weed.
- IoT environmental monitoring: Sensors record temperature, humidity, light intensity, rainfall and soil moisture. GPS, Wi-Fi/4G connectivity, a remote-observation camera and a 7-inch HDMI display support monitoring and operation.
- Field workflow and energy: The robot surveys the cultivation area, builds a movement map, works through the field by grid cells and updates a digital map. It then returns to its home position. Batteries combined with a solar panel supply the power.
IMPACT
The robot is designed to reduce manual work and chemical use in vegetable farming, and to give farmers data on field conditions, weeds and signs of disease. Weeds are treated one by one with a laser rather than sprayed. Trial operation was planned for July to early August 2026, and the final reporting and refinement stage for 1 to 30 August 2026.



