Missile Guidance Law Based on Robust Model Predictive Control Using Neural-Network Optimization.

Area of Science:

• Aerospace Engineering
• Control Systems Theory
• Applied Mathematics

Background:

• Missile interception requires advanced guidance systems to counter dynamic target maneuvers.
• Existing methods may struggle with real-time adaptation to unpredictable target accelerations.

Purpose of the Study:

• To develop a robust guidance law for missile interception using model-based predictive control.
• To address target acceleration as a bounded disturbance within the control framework.

Main Methods:

• A novel guidance law based on model predictive control (MPC) is formulated.
• The MPC problem is converted into a constrained quadratic programming (QP) problem.
• A linear variational inequality-based primal-dual neural network solves the QP problem over a receding horizon.

Main Results:

• The proposed guidance law effectively incorporates missile-inside constraints.
• Online solutions to parametric QP problems enable real-time optimal control decisions.
• Simulation studies demonstrate the guidance control law's effectiveness and performance.

Conclusions:

• The developed model-based predictive control guidance law offers a robust solution for missile interception.
• The neural network-based QP solver facilitates real-time constrained optimal control.
• The approach shows significant promise for enhancing missile defense capabilities.