Measurement of Motion Characteristics Using Neural Networks

A neural-network-based algorithm is offered for measuring the motion characteristics of a landing module. The measurement system is based on recording photon scattering. It uses a central photon source and four photon detectors for measuring the altitude, the traverse speed and the slope angle of the landing module. The model is intended for the initial stage of developing the information measuring system. The model allows to estimate the results of altitude, traverse speed and slope angle measurement, analyze the geometrical arrangement of photon sources and photon detectors, investigate the influence of chemical composition of underlying surface. The algorithm uses the information obtained from the photon detectors arranged on the lunar landing module and implements a state-space model of the landing module. A computer model of module motion produces training data. The generalization results allow to estimate measurement uncertainty. The simulation results of the algorithm and the estimated error of measurement of height, speed and angle of descent vehicle are presented.