Guidance And Control Of Agile Missiles Subject To Input Saturation

In order to improve the hit to kill intercept probability against highly maneuverable targets, advanced missiles employ multiple actuators to enhance maneuverability and agility. Actuators employed in these missiles include aerodynamic control surfaces and reaction jets. In this thesis, we study the guidance and control problems of agile missiles with dual aerodynamic control surfaces and reaction jets subject to input saturation.Based on variable structure control with sliding sector and finite time stability theory, a novel finite time sliding sector is proposed. For the three dimensional nonlinear model of guidance system, the analytical solution of a state dependant differential Riccati equation is achieved to determine the finite time sliding sector. With the assumption that the upper bound of the target acceleration can be estimated a priori and taking the tangential relative velocities as system states, a finite time sliding sector guidance law is designed based on the proposed sliding sector. Under the finite time silding sector guidance law, the system states can move into sliding sector from any initial states outside the sector. Inside the sliding sector, the tangential relative velocities converge to zero in finite time which ensures the missile achieve direct collision with the target. In the presence of acceleration saturation, the proposed guidance law can ensure the tangential relative velocities converge to zero before the end of the guidance process under specific initial conditions.In practical applications, the target acceleration and its upper bound are difficult to be accurately measured without delay and the autopilot lag of intercept missiles usually cause bad influence on the guidance precision. Considering the target acceleration as bounded uncertainty, the three-dimensional nonlinear guidance law with missile autopilot dynamics is designed by using dynamic surface control. In the design process, the upper bound of target acceleration does not need to estimated a priori. Instead, a parameter update law is designed by Lyapunov stability theroy to estimate the upper bound online. An extra auxiliary low-pass filter is introduced to compensate for the effect of input saturation. In the expression of the adaptive dynamic surface guidance law, the high-order derivatives of the line-of-sight angle does not occur, which means that the guidance law is easy to be implemented. The relations and differences between the proposed guidance law and proportional navigation, sliding mode guidance law, H∞ guidance law are discussed. Numerical simulation results shows that the proposed guidance law is able to compensate for the autopilot lag and the effect of input saturation. It can provide a high guidance precision on intercepting a maneuvering target.The missile is controlled by the combination of aerodynamic control surfaces and reaction jets. First, the reaction jets are neglected in the design of aerodynamic control sufaces control system. Then, the reaction jets control system is designed based on the aerodynamic control surfaces control system. In the presence of aerodynamic control surfaces saturation, two approaches are developed to design the controller by utilizing the invariant set analysis with disturbance rejection. One is the control input is permitted to be saturated, and the other is the control input never reaches the saturation limits. Numerical simulation shows that both of the approaches can ensure the missile acceleration tracking an acceleration command fast. The two approaches are robust against the interactions between the airflow and the reaction jets which verifies the robustness of the proposed approaches. Compared with the approach which the control input does not reach the saturation, the other approach which the control input is permitted to be saturated comsumes less reaction jets, because the aerodynamic control surfaces are fully used when tracking the acceleration command.For a control system with multiple actuators, control system design without actuator control distribution sometimes can result in control efforts counteracting with each other. Control allocation is an effective method to deal with the control system design of multiple actuators. Base on the analysis of the roles of aerodynamic control surfaces and reaction jets in buiding the angle-of-attack, it is shown by simulation results that autopilot design without control allocation result in reduced missile controllability and efficiency. The aerodynamic control surfaces counteract with reaction jets, they can not fully develop their actions. There are two steps in the autopilot design. The virtual control law is designed by utilizing adaptive dynamic surface control first. Then, the virtual control effort is distributed to individual actuator settings by dynamic control allocation. Compared with pseudoinverse control allocation, the aerodynamic control surfaces under dynamic control allocation do not reach the saturation limits with rational allocation. In the tracking process, aerodynamic control surfaces cooperate with reaction jets, and the two actuators fully developed their actions.