Cooperative attack decision for BVR air combat based on neural network

In order to solve the problem of beyond visual range (BVR) formation cooperative attack, a muti-objective distributed neural network decision model was established. Based on the situation assessment function, the advantage and disadvantage regions in BVR air combat were divided. Moreover, four air combat patterns for the specific regions, which could be taken as the training samples for the GA-BP situation assessment neural network, were proposed. Therefore, the situation advantage matrix of multi-aircraft cooperative air combat was obtained. According to the situation assessment matrix, the effective attack order could be obtained with the Hopfield neural network. A concept of second round attack was introduced, which could maximize the superiority of attack situation to the enemy attack and minimize the threat level of enemy attack. The simulated results of BVR multi-aircraft cooperative attack show that after analyzing each index for air combat, the proposed model can provide the situation assessment indexes for both sides and effective attack order.