PSO Tuned Non-Linear Fuzzy PID Trajectory Control of Row Seed Sowing Autonomous Agricultural Robot

Abstract

Agriculture is considered one of the most important economic activities in Ethiopia. About 85% of the Ethiopian population are dependent on agriculture directly or indirectly through a traditional farming mechanism. Seed sowing using the human hand is a highly inefficient process that requires a lot of human effort, which leads to health concerns and is a time consuming as well as tedious activity for farmers. This thesis describes the development of the Row Seed Sowing Autonomous Agricultural Robot, which is an automated system of differential derive mobile robot for sowing the seed when the robot strictly follows the generated trajectory of the farmland to minimize the working cost, time for digging, sowing, and increase the agricultural productivity. The mathematical model of the nonholonomic differential derive mobile robot involves two identical series of DC Motors for deriving, robot dynamics and kinematics, sensors that provide information about the trajectory of the farmland according to the generated trajectory. Tracking errors where the robot sways off its path with one or more wheels lead to sowing the seed unproper out of the required trajectory on the farmland. The reference trajectory tracking nonlinearity and adjustments of the DC motor speed are both overcome using the controller for the motion model of the differential derive mobile robot that operates on different paths. This means controlling the velocity of the front two motors independently to control the two-wheel velocities. So, a PSO algorithm tuning of a fuzzy PID controller is proposed. The PID gains, the number of fuzzy control rules, the center, and the widths of the Gaussian membership functions are all parameters to be determined simultaneously to track the path and minimize tracking errors to sow the seeds properly. Tests are also carried out for different trajectories to evaluate the performance of the controllers. PSO FPID, FPID, and PID controller position and orientation tracking ability, tracking error of reference, and desired trajectory are also evaluated, and compared the POS tracking error as well as integral performance criteria of IAE and ISE with time through simulation studies in MATLAB/Simulink for different paths. The results of the position and orientation tracking error of the straight-line trajectory tracking for PSO FPID, FPID and PID are (0.525, 1.529, 2.992) and (-0.276, -0.515, -1.136) respectively. The result shows that the PSO-tuned FPID trajectory controller has better performance than fuzzy PID and conventional PID controllers