Model Öngörülü Kontrole Dayalı Hava Savunma Füzesi için Karışık Tamsayılı Karesel Programlamalı Otopilot Tasarımı
Özet
In modern combat scenarios, the effectiveness of tactical missiles relies heavily on their maneuverability and control systems. Enhancing missile performance, considering the challenges posed by highly maneuverable or high-speed ballistic missile targets and the need for faster response times, depends on actuator capability. To address the demand for enhanced maneuverability, in addition to aerodynamic surfaces, secondary actuators such as lateral thruster systems are explored. Unlike traditional control mechanisms, these systems utilize reaction forces from ejected high-velocity gases, offering consistent control efficacy across all flight phases, especially in low dynamic pressure conditions where aerodynamic surfaces may be less effective. However, the on-off nature and limited number of propulsion systems necessitate defining constraints in controller design. In addition to lateral thruster constraints, aerodynamic surfaces have limitations of angle and angular rate.
Within the scope of the thesis, aiming to provide an optimal solution under defining constraints, an approach for Mixed Integer Quadratic Programming (MIQP) based on Model Predictive Control (MPC) is proposed. In MPC, a dynamic model of the system is used to forecast its future behavior by considering its current state and the control inputs applied. MIQP is an optimization problem characterized by a quadratic objective function and decision variables that must be integers. MIQP is employed because the lateral thruster is treated as an integer variable. In the study, the aerodynamic fin angle constraint, which is another system constraint, has been defined as an inequality constraint. However, the system struggles with stability when the fin constraint is active while there is already an equality constraint in the system. Therefore, one of the aims of this thesis is to update the value of the fin in the penalty function to ensure that the fin constraint does not become active. This is attempted using an approach called the "MPC-Based Adaptive Weight Estimation Algorithm" method. The method simply aims to reduce the effect of the fin by increasing the weight if the fin limit is active, checking narrow windows from the MPC solution to see whether the fin limit is triggered.
The main target of the autopilot (MIQP-MPC) is to track the acceleration command that comes from guidance. The autopilot is designed for only the pitch channel of a missile in the linear domain, and its performance is analyzed for different scenarios. Moreover, the robustness of the autopilot against the jet interaction effect between lateral thrusters and the air stream is also examined. This interaction in missile systems is a critical aspect of aerodynamic control and creates complex aerodynamic forces and moments, which can alter the missile's flight path. The expelled gas jets disrupt the smooth flow of air around the missile, potentially causing turbulence and affecting overall stability and control. Finally, to compare the performance of the autopilot integrated with the lateral thruster with the conventional fin-controlled autopilot structure and to evaluate the designed autopilot in the guidance loop, a ballistic missile target scenario was created and the results were evaluated.
