<![CDATA[The output power of photovoltaic systems (PV) can be significantly reduced by dust accumulation. Among various cleaning methods, robotic cleaning is currently the most popular choice because it minimizes human effort and reduces the risk of damaging PV cells. However cleaning robots can be impacted by various external disturbances, including wind, rain, lightning, snow, , and vibrations. Additionally, sensor errors related to slip, position, velocity, acceleration, and varying electrical parameters can also affect their performance.Several methods have been proposed in the literature for tracking the robotic cleaning trajectory of PV systems. Nevertheless, most of these methods struggle in the presence of disturbances and often have prolonged convergence times. This paper aims to propose a Model Reference Adaptive Control system to maintain optimal performance and extend the lifespan of PV panels, minimize power losses, reduce convergence time, achieve optimal tracking of the desired cleaning trajectory amidst disturbances, and decrease the dependence on multiple sensors. In our study, we utililized the iRobot solar panel developed by Aravind et al., which has a power capacity of 250 W and weighs 250 kg. This iRobot can effectively clean approximately 930 solar panels of the Kyocera Solar KC 130 GT module, which measures 1.425 m in length and 0.652 m in width. The iRobot operates for 4 hours, covering an area of 864 m², and can clean a surface area of 0.06 m² in one second. We conducted simulations using the proposed MRAC algorithm in Matlab/Simulink software, comparing the results with those obtained from a Proportional Integral Derivative (PID) algorithm. The results demonstrate that the MRAC approach achieves a shorter convergence time and greater precision in following the desired cleaning trajectory of the robot, even in the presence of disturbances, compared to the PID algorithm.]]>
