Deep Neural Network Based Cooperative Guidance Law for Speed-Varying Interceptors

This paper studies the problem of three-dimensional (3-D) cooperative guidance for speed-varying in-terceptors. A 3-D cooperative guidance strategy is designed by adding a cooperative term to proportional navigation guidance (PNG). First, a deep neural network (DNN) is constructed for predicting the time-to-go of the speed-varying interceptor under PNG. Then, a cooperative term is derived on the basis of predicted time-to-go information. Moreover, the time-to-go consensus error system is proven to be input-to-state stable (ISS) under the designed guidance law. Finally, numerical simulations are conducted to illustrate the validity of the proposed method.