This thesis introduces the new guidance algorithm of BTT (Bank-To-Turn) missiles using the neural networks. In general, neglected the dynamics of BTT missile, the PN(Proportional Navigation) guidance provides a good interception performance. This algorithm has it for its object to obtain the new guidance law which makes the true acceleration and roll angle of BTT missile be at one with PN guidance law using the inverse dynamics achieved by neural networks.
The applications of this guidance loop are accomplished for two BTT models. One is the first order acceleration and roll delay model including an autopilot, the other is the nonlinear full dynamics model without an autopilot. After this, the performance of new guidance algorithm for each case is compared with that of the existing PN guidance.