Application of Fuzzy Logic Methods to Develop a Quadcopter Landing Control System

The article addresses a pressing scientific and applied problem related to the automation of the landing process of unmanned aerial vehicles, in particular quadcopters, which today occupy a leading position among the various types of drones and are widely employed in both military and civilian domains. The landing stage constitutes an obligatory phase in the operation of such aircraft, yet at the same time it is one of the most complex, since in its final stage the vehicle must ensure safe contact with a solid surface. Consequently, the development of an effective algorithm and system for automatic quadcopter landing becomes the central objective of the conducted research. Existing landing control solutions based on ultrasonic, infrared, thermal, or radio signals require the use of additional equipment, which complicates their integration into small and medium-sized aerial platforms. In this context, the most rational approach is considered to be the use of the quadcopter’s onboard camera in combination with a specially designed landing marker placed on the platform, constructed from simple geometric figures that are easily recognizable from altitude. To achieve the stated objective, a mathematical model of quadcopter motion was created in the Matlab Simulink environment. This model takes into account the thrust of each of the four propellers, thereby enabling the reproduction of the aircraft’s displacement along the coordinate axes and the simulation of landing scenarios. The automatic control system was implemented using a fuzzy controller, which operates on the basis of three input variables representing the deviations of the current coordinates from the target landing point along the x, y, and z axes. The outputs are expressed as thrust forces for each motor. A total of thirty fuzzy inference rules were constructed, ensuring adequate responsiveness to deviations from the designated trajectory and altitude. Experimental investigations conducted in the modeling environment confirmed the effectiveness of the proposed solution. All tests demonstrated stable guidance of the quadcopter toward the designated marker and successful completion of the landing process. The average value of the static error after landing corresponds to the requirements of the technical specification and is regarded as acceptable for automated landing systems designed for platforms with limited surface area. The dynamics of the process were characterized by the absence of significant oscillations and overshoot, a smooth reduction of speed in the final phase, and reliable attainment of the designated point. The obtained results indicate that the proposed approach possesses substantial practical value and may be effectively employed for the creation of real-world quadcopter landing control systems.