In the chapter, the author considers the possibility of applying modern IT technologies to implement information processing algorithms in UAV motion control system. Filtration of coordinates and motion parameters of objects under a priori uncertainty is carried out using nonlinear adaptive filters: Kalman and Bayesian filters. The author considers numerical methods for digital implementation of nonlinear filters based on the convolution of functions, the possibilities of neural networks and fuzzy logic for solving the problems of tracking UAV objects (or missiles), the math model of dynamics, the features of the practical implementation of state estimation algorithms in the frame of added additional degrees of freedom. The considered algorithms are oriented on solving the problems in real time using parallel and cloud computing.
TopAdvanced Kalman Filter
The extended Kalman filter uses the model of a stochastic continuous system:
𝑥˙ = (𝑥, 𝑡) + 𝑤
(1)and discrete measurements:
z
k = h
k(x(t
k)) + y
k(2)where 𝑥 is the state vector of the system, 𝑤 is the noise of the system, y𝑘 is the measurement noise.
The task of filtering is to find a measurement function 𝑧𝑘 an estimate of the state vector of the 𝑥’𝑘 system that minimizes the variance of the error x'k-x (tk).
Let an estimate of the state vector 𝑥’𝑘 − 1 be obtained at time 𝑡𝑘 − 1. Based on this estimate, the forecast of the state vector estimate 𝑥’𝑘 ’(a priori estimate) is built, then zk measurements and correction of the estimate a priori based on x’k’ ’measurement results (a posteriori estimate) are carried out (Wilson, R. C. & Finkel, L., 2009). An estimate a priori of the state vector x’k ’is calculated by integrating the model equation (3):
with initial conditions 𝑥’(𝑡) = 𝑥’k-1, 𝑥’(0) = 𝑥0.
The a priori estimate for the covariance error matrix for the linearized equations in 𝑃’k increments is calculated as (4),(5), (6):
𝑃
k’= φ𝑃
k’’-1φ
T +𝑄,φ = 𝐸 + 𝐹δ𝑡
With the initial conditions 𝑃’(𝑡) = 𝑃𝑘’’−1, 𝑃(0) = 𝑃0. Where, 𝑄 is the covariance matrix of the noise of the system (Doucet, A. & Johansen, A., 2009). A posteriori estimate for the state vector and covariance matrix errors are constructed as follows: