Projectile interception using a neural network

Neural networks (NNs) serve as a versatile and robust tool for many engineering problems. A NN can be used as a very accurate approximator for multivariable nonlinear equations such as that of the projectile interception. While there is an established algorithm to solve the equation of the projectile interception, it is computationally time consuming for real time applications. The accurate and reliable solution given by a NN reduces the computation time to a constant low value. In this paper, a NN is used to develop a real time projectile interception controller with a performance comparison against the conventional method of using an iterative numerical algorithm. In this study it is demonstrated that the NNs are capable of yielding comparable accuracy and reliability while being less computationally time consuming.