Three-Dimensional Resilient Cooperative Guidance Under Varying Speed and Field-of-View Constraint

This article studies the challenging problem of field-of-view constrained cooperative guidance, considering variations in missile speeds and the influence of abnormal missiles. To address the issue of real-time prediction of the time-to-go for speed-varying missiles, a deep neural network is built and trained, whose input is a five-dimensional feature vector skillfully selected from the flight states. Then, by taking the form of proportional navigation guidance including a time-varying gain, a three-dimensional (3-D) cooperative guidance law is devised utilizing local time-to-go information. Moreover, to resist the influence of abnormal missiles, a resilient cooperative guidance law with the similar structure is further designed by virtue of a time-to-go sequence reduction method. The lead angles of missiles guided by the designed cooperative guidance laws can be confined within an allowable range, thus avoiding losing the target. In addition, detailed analysis is conducted on the convergence of the designed guidance laws. A series of numerical simulations are executed for validating the effectiveness of the designed guidance laws.