Browsing by Author "Fragulis, George F."

Filter results by typing the first few letters
Now showing 1 - 2 of 2
  • No Thumbnail Available
    Item
    A New 3D Sliding Pursuit Guidance Law for Fixed Wing Combat Drone Piloting: Application to El-Djazaïr 54
    (World Scientific, 2025) Bekhiti, Belkacem; Fragulis, George F.; Hariche, Kamel
    This paper introduces a new finite-time neural adaptive nonlinear 3D sliding pursuit guidance law designed for autonomous control of fixed-wing Unmanned Aerial Vehicles (UAVs) targeting a maneuvering object. The central innovation in the control strategy is the incorporation of sliding control in pure pursuit, which significantly enhances robustness against uncertainties and variations. Simulations were conducted using a specific combat drone model (El-Djazaïr 54), within a real-time virtual Simulation Platform for Aircraft Control System (SP-ACS). The control approach is model-based, with an initial identification phase before testing and validation. To identify unknown, variable, and classified aerodynamic parameters, the Total Least Squares Estimation (TLSE) method was employed. The mean values of aerodynamic coefficients were calculated, with any deviations treated as modeling uncertainties to be managed by the robust control law. Simulation results demonstrate that the El-Djazaïr 54 drone exhibits excellent performance in tracking the moving target and maintaining robustness despite modeling uncertainties
  • No Thumbnail Available
    Item
    On The Block Decomposition and Spectral Factors of λ -Matrices
    (Arxiv, 2018) Bekhiti, Belkacem; Dahimene, Abdelhakim; Hariche, Kamel; Fragulis, George F.
    In this paper we factorize matrix polynomials into a complete set of spectral factors using a new design algorithm and we provide a complete set of block roots (solvents). The procedure is an extension of the (scalar) Horner method for the computation of the block roots of matrix polynomials. The Block-Horner method brings an iterative nature, faster convergence, nested programmable scheme, needless of any prior knowledge of the matrix polynomial. In order to avoid the initial guess method we proposed a combination of two computational procedures. First we start giving the right Block-QD (Quotient Difference) algorithm for spectral decomposition and matrix polynomial factorization. Then the construction of new block Horner algorithm for extracting the complete set of spectral factors is given.