Author：Peng Yen-Chia/Manager, Industrial Technology Research Institute
Wu Sung-Yi/Associate Engineer, Industrial Technology Research Institute
Chang Chia-Wei/Assistant Researcher, Taichung District Agricultural Research and Extension Station

In an era of rapid technological advancement, agriculture— a sector fundamental to daily human life—is undergoing profound transformation. Amid escalating challenges such as population aging, climate change, and labor shortages, agricultural automation and smart farming have become essential to ensuring food security and sustainable development. Within this wave of change, agricultural robotics plays a pivotal role. The evolution from early-stage remote-controlled machines to the latest technology of self-propelled systems not only exemplifies the rapid progress of technology but also marks a fundamental shift in agricultural production methods.

Traditionally, Farmers endured long hours of intensive labor in the fields, facing high workloads in relatively inefficient working conditions. While the advent of agricultural mechanization—with tools such as cultivators, sprayers, tractors, and combine harvesters—greatly enhanced productivity, human operation remained indispensable. In recent years, the concept of automated agriculture has gained significant traction, and agricultural operations are moving steadily toward greater precision and intelligence. Modern farm operators are gradually transforming into “agricultural engineers,” making management decisions through data analysis and remote monitoring technologies. By using smart devices to track crop growth, soil conditions, and climate variability, they are achieving more scientific and data-driven agricultural management.

The agricultural robotics team of the Industrial Technology Research Institute (ITRI), through a cross-disciplinary collaboration initiative, has developed a robotic platform for field tasks, which has already been piloted in vineyard environments for sprayer development. As agricultural vehicles advance toward full autonomy, sensor technologies have emerged as a critical factor. New-generation technologies have endowed traditional agricultural machinery with the equivalent of "eyes," "ears," and a "brain," enabling these machines to perceive their surroundings, make real-time decisions, and perform complex operations. For example, GPS receivers combined with Real-Time Kinematic (RTK) positioning systems allow centimeter-level accuracy in navigation, enabling agricultural robots to follow predefined routes. Inertial Measurement Units (IMUs) provide attitude and motion data, which are especially vital in areas with complex terrain or unstable GPS signals. Light Detection and Ranging (LiDAR) equips agricultural machines with 3D environmental perception capabilities, allowing the creation of spatial maps of their surroundings—an essential feature for safe navigation in orchards or uneven terrain. Advanced imaging systems, including RGB and multispectral cameras, significantly enhance visual recognition. The former cameras are used for obstacle detection, while the latter assess crop health and maturity levels. Ultrasonic sensors act as the "tactile sense" of the machinery, detecting nearby obstacles and supplying critical data to control systems for collision avoidance or emergency stops.

The integration of sensing technologies has made autonomous agricultural vehicles a reality—not merely as mechanical tools, but as intelligent operational systems with full environmental awareness. In practical applications, these technologies enable precision agricultural tasks. For instance, during pesticide application, crop health data from imaging systems can be analyzed in real time. When combined with information from soil sensor data, the system can dynamically adjust the dosage and placement of agrochemicals, significantly improving resource efficiency, reducing waste, and promoting healthier crop growth.

However, transforming raw sensor data into real-time decisions requires powerful computing capabilities, bringing edge computing to the forefront of development. Edge computing replaces reliance on cloud processing by enabling data analysis to be performed locally on the machine itself. This allows for immediate responses: for example, if sensors detect an obstacle ahead, traditional cloud-based systems may experience delays due to data transmission, whereas edge computing allows the vehicle to compute an evasive path within milliseconds, ensuring uninterrupted and safe operation. In remote agricultural regions with limited or unstable internet connectivity, edge computing becomes especially vital. It allows autonomous machines to function efficiently even without cloud access, improving operational reliability and adaptability while reducing dependence on network bandwidth.

Although the development of agricultural automation draws heavily from experience in the industrial sector, it faces far more complex challenges. Unlike automated guided vehicles (AGVs) and autonomous mobile robots (AMRs) commonly used in controlled factory environments, agricultural machinery must operate in unpredictable outdoor conditions. These include rugged and uneven terrain, rapidly changing weather, and a wide range of potential obstacles in the field—conditions that demand superior environmental perception and adaptability. Nevertheless, autonomous agricultural systems can still benefit greatly from industrial automation expertise. Valuable insights include precise path planning, multi-machine coordination and scheduling systems, intuitive human-machine interfaces (HMIs), and modular design principles that enhance scalability and maintenance efficiency.

Looking forward, autonomous agricultural machinery will continue to evolve toward greater intelligence, efficiency, and reliability. This progress marks not only a major milestone in agricultural automation, but also signals a transformative shift in the very model of agricultural production. In the near future, intelligent field robots will perform autonomous tasks with unprecedented precision and productivity. These advancements will significantly enhance the working conditions of farmers, reduce physical labor, and accelerate the transition toward more sustainable and environmentally friendly farming practices. This technological revolution will not only change how we plant and harvest crops, but also redefine the relationship between humans and the land, opening up new possibilities for food security and sustainable agricultural development.