A method for predicting the flight movement of the reference points in the drone formation
In order to achieve accurate prediction of the flight movement of the lead aircraft/virtual reference point in the drone formation,a method for estimating the acceleration of the lead aircraft/virtual reference point based on RBF neural network is proposed.The updating method of the actual weights of the neural network is designed,and the improved intelligent waterdrop algo-rithm is used to obtain the Gaussian radial basis function centers(c1 max,c2 max)of the neural network.In the above algorithm,the ve-locity and acceleration of waterdrops simulate the velocity and acceleration of drones.According to the norms of matrices composed of position deviations and velocity deviations of waterdrops,as well as soil content of grid,the grid position of waterdrops is up-dated.When the computational stop condition is met,(c1 max,c2 max)at the grid with the least soil content is output as the center of basis function.Simulations show that the drone formation can quickly respond to maneuvering flight,and the position and velocity deviations of each aircraft after formation transformation are small enough.It can be seen that the acceleration estimation of the lead aircraft/virtual reference point is accurate enough,and the estimation method is effective.
万方数据
维普
