Harvest-Mate: Automation Meets Agriculture

The palm oil industry is a vital economic driver in Malaysia, yet it faces significant challenges in labour shortages, productivity, and sustainability. Despite technological advancements in agriculture, harvesting Fresh Fruit Bunches (FFB) from oil palm plantations remains heavily reliant on manual labour, presenting risks to worker safety and efficiency. Addressing these issues, a team of third-year engineering students from Universiti Malaya has developed an innovative solution — Harvest-Mate, an automated palm oil harvesting machine designed to transform the industry.

The idea for their brainchild, Harvest-Mate, emerged from the students’ Integrated Design Project (IDP), which challenges participants to solve real-world problems. The team observed the inefficiencies and hazards of manual FFB harvesting, particularly the physical strain on labourers and the dependency on human precision for identifying ripe fruits.

"Our goal was to automate the harvesting process to reduce reliance on manual labor while improving efficiency and worker safety," explained the team.

Harvest-Mate is a testament to the integration of cutting-edge technologies with practical applications in agriculture. The machine is equipped with a tracked base to navigate the often uneven terrains of palm oil plantations. Its robotic arm, fitted with a motorised jigsaw blade, performs the crucial task of cutting FFBs precisely.

Key Features Include:

• Real-Time Detection and Segmentation: Using machine learning-powered YOLOv8, Harvest-Mate identifies and segments ripe FFBs from the surrounding vegetation. The system is trained with datasets tailored to plantation conditions, ensuring high accuracy.

• Robotic Arm with Vibrational Cutting: The arm, guided by depth cameras and IMU sensors, uses vibration-enhanced blades to make efficient cuts, reducing mechanical strain and improving speed.

• Sensor Suite for Navigation: GPS, IMU, and Lidar sensors provide spatial awareness, enabling smooth navigation and obstacle avoidance.

• Energy-Efficient Design: The machine is powered by modular lithium-ion batteries. It can harvest 230 FFBs in 16 hours while consuming 1093.44 watt-hours of energy.

The design employs a modular and standardised approach, allowing easy mechanical, electronic, and software integration through the Robot Operating System (ROS). This system mimics natural communication networks, ensuring seamless interaction between sensors, actuators, and controllers.

Harvest-Mate demonstrates a 123.3% increase in productivity compared to manual harvesting. While labourers typically harvest 103 FFBs in 8 hours, Harvest-Mate achieves 230 FFBs in 16 hours. This improvement addresses the industry’s critical labour shortage and significantly enhances operational efficiency. Moreover, the machine eliminates the need for labourers to climb trees or work in hazardous conditions, reducing the risk of accidents and improving overall safety.

The development of Harvest-Mate was not without obstacles. The team faced difficulties optimising the robotic arm’s cutting mechanism, ensuring precise navigation across uneven terrain, and calibrating sensors for real-time detection. "We conducted extensive testing, from analysing cutting angles and leaf types to assessing sensor accuracy, to ensure Harvest-Mate meets the industry’s demands," the team shared.

The integration of YOLOv8 for detecting and segmenting FFBs proved to be a game-changer, providing the accuracy needed for autonomous operations. Biophilic design principles further enhance the machine's ability to operate in harmony with the natural environment, minimising disruption to the ecosystem.

Harvest-Mate is a transformative innovation for the palm oil industry, offering solutions to long-standing challenges:

• Economic Efficiency: By automating labour-intensive tasks, the machine reduces costs and enhances profitability.

• Sustainability: Precision harvesting minimises waste and collects only ripe FFBs, maximizing oil yield.

• Environmental Benefits: The electric-powered design reduces carbon emissions, aligning with global sustainability goals.

• Workforce Safety: Automated systems eliminate the need for labourers to perform high-risk activities, improving workplace safety standards.

The team envisions scaling Harvest-Mate for widespread adoption in the palm oil industry. Future upgrades may include enhanced battery systems for longer operational hours, advanced cutting technologies, and expanded AI capabilities for multi-tasking in plantations. "We aim to collaborate with industry stakeholders to refine and commercialise Harvest-Mate, ensuring it becomes a staple in modern agriculture," said the team. The integration of robotics and machine learning in agriculture, as demonstrated by Harvest-Mate, represents a significant step toward achieving sustainable and technologically advanced farming practices.

Harvest-Mate is not just a machine—it is a vision of what the future of agriculture could be: efficient, sustainable, and safe. This groundbreaking project exemplifies the potential of student innovation to address global challenges and improve industries.

Field testing with Harvest-Mate in palm oil plantation

Award Giving Ceremony of Engenius'24 Integrated Design Project Conference

Close-up of Harvest-mate in action

From left: Cheah Yih Xiang, Khaw Thong Xiang, Assoc. Prof. Ir. Dr. Anis Salwa, Soong Peng Xiang, Sherrie Goh

Picture taken during online interview with the research team

#SustainableAgriculture #AgriTech #PalmOilIndustry #InnovationInFarming #Automation #SmartFarming #GreenTech #FutureOfAgriculture #EcoFriendly #PrecisionHarvesting #AIinAgriculture #SustainableFuture #TechForGood #StudentInnovation #AgriEngineering

---

Researcher featured:

Ir. Dr. Wong Wei Ru, Assoc. Prof. Ir. Dr. Anis Salwa Mohd Khairuddin, Prof. Ir. Dr. Yap Hwa Jen, Assoc. Prof. Ir. Dr. Chuah Joon Huang & final-year electrical engineering students, Soong Peng Xiang, Khaw Thong Xiang, Sherrie Goh Xun Yi, and Cheah Yih Xiang

Ir Dr Wong Wei Ru

Department of Electrical Engineering

Faculty of Engineering, Universiti Malaya

https://umexpert.um.edu.my/weiru

For inquiries, please contact:

E: weiruum.edu.my

T: 03-79672659

https://www.scopus.com/authid/detail.uri?authorId=55550442400

Author:

Ms Puungkodi Paramasivam

Puungkodi Paramasivam is a fast-track PhD candidate at the Faculty of Languages and Linguistics, Universiti Malaya, who finds solace and purpose in writing amidst the challenges of academic research. This is one of her many contributions to the UM Research Bulletin, a platform she values for showcasing groundbreaking work while allowing her to channel her creativity. Writing, she says, serves as a positive distraction from the demands of her PhD journey, enabling her to engage meaningfully with diverse fields of study. Proud to be an author under the UM Research Bulletin, Puungkodi continues to highlight innovative research and inspire others with her thoughtful narratives.

Copyedit:

Siti Farhana Bajunid Shakeeb Arsalaan Bajunid, Assistant Registrar, UM

Nurhazrin Zanzabir, Assistant Administrative Officer, UM