Automated Weeding Robot
A GPS-guided field robot that navigates crop rows and removes weeds mechanically.
- Role
- Co-author (4th of 6)
- Team
- 6 students — IUB
- Timeline
- Sep 2025 – Dec 2025
- Organisation
- Independent University, Bangladesh
- Code
- Private repository
- Reading time
- 1 min
Overview
Manual weeding is slow and herbicides are expensive and harmful. The robot drives between crop rows and cuts anything growing in the gap — a deliberately simple, non-vision approach that works without costly cameras or models.
The problem
- Navigate rows reliably outdoors.
- Avoid obstacles and people.
- Remove weeds without chemicals or expensive machine vision.
Context & constraints
Weed detection is positional: any plant between the rows is treated as a weed.
Approach
- Control — Arduino Mega (ATmega2560) running the navigation loop, with ESP8266 Wi-Fi for telemetry.
- Sensing — four HC-SR04 ultrasonic sensors (2–400 cm) and a NEO-6M GPS for waypoints.
- Drive — four 300 RPM gear motors with differential steering through BTS7960 drivers at 10 kHz PWM.
- Cutting — a 775 motor at about 21,000 RPM.
- Power — a 12 V, 2200 mAh lithium-ion pack for around four hours of operation.
Key capabilities
- GPS waypoint navigation (±2.5 m)
- Obstacle avoidance under 0.5 s
- Mechanical, chemical-free weeding
- ≈4 h battery life
Architecture
Sensing
Control
Actuation
A classic sense → decide → act loop on a microcontroller. Select a component to trace its connections.
Quality & delivery
Testing
Lab and field trials measured obstacle range, GPS accuracy, avoidance delay, battery life and weed-removal rate.
Outcome
- Weed removal (lab)
- 85%
- 75% in field trials
- GPS accuracy
- ±2.5 m
- NEO-6M
- Obstacle avoidance delay
- <0.5 s
- Battery life
- ≈4 h
- 12 V, 2200 mAh
A working prototype documented in an IEEE-style paper with lab and field results.
Lessons
- Field conditions take a real bite out of lab numbers — plan for the gap.
Building something similar?
Tell me what you’re working on — I reply to every serious message.
Related work
Gesture-Based Messaging for Assistive Care
Two ESP32 units over Bluetooth Classic: a wearable sender with an MPU9250 motion sensor and a receiver with an OLED and buzzer. Reliable at 10–20 metres.
- C / C++
- ESP32