Reactive fuzzy control and field validation of a low-cost autonomous tracked robot for plantain crop row following

Abstract This study details the design, development, and preliminary field validation of a low-cost tracked mobile robot intended for autonomous navigation within plantain crop rows. Plantain is a significant crop in numerous tropical regions; however, robotic solutions for operations beneath plantain canopies are limited, presenting a compelling opportunity to explore the application of mobile robots in agriculture. In this regard, low-cost under-canopy robots offer promising alternatives for facilitating frequent in-field supervision and data collection. The proposed platform integrates two 12V DC motors with encoders, an ESP32 microcontroller for low-level proportional-integral (PI) motor control, and a Raspberry Pi 5 running Robot Operating System 2 (ROS 2) for high-level control. The perception system incorporates a YDLIDARX2 LiDAR scanner and a BNO055 inertial measurement unit (IMU) to estimate both the lateral distance to the crop row and heading orientation. These measurements were used by a fuzzy logic controller to generate differential angular velocity commands, maintaining a target distance of 1.0 m from the right-hand crop row. Field experiments conducted under a canopy and on uneven, muddy terrain demonstrated reliable motion and acceptable lateral error stabilities. The integration of IMU-based heading estimation with the fuzzy reactive controller enhanced path alignment and effectively managed the external disturbances. The results suggest that the combination of low-cost components with a reactive fuzzy control strategy facilitates reliable under-canopy navigation in plantain plantations and establishes a foundation for the future popularization of low-cost robots supporting agricultural tasks in plantain crops.