Computational Guidance Law Based on Data- Driven Online Adaptive Critic Designs

This paper presents an adaptive critical design method, which using neural network (NN) approximator in Actor/Critical structure to calculate the approximate solution of the optimal control problem of missile-target interception scenario online. The method is based on Policy Iteration (PI) algorithm, the PI algorithm starts from the initial admissible control to find the optimal solution iteratively. In this paper, the initial admissible control input is chosen as proportional navigation control law. By online data sampling, the implementation of the algorithm does not need to use the explicit knowledge of the internal dynamics of the nonlinear system. The numerical simulation is provided to validate the effectiveness of the proposed guidance law.