Rescue orbit decision-making method based on RBFNN under rocket thrust descent fault

Abstract

The invention discloses a rescue orbit decision-making method based on RBFNN under a rocket thrust descent fault. The method comprises the following steps: building a secondary flight kinetic equationof an ascending section of a rocket in a geocentric inertial coordinate system, and building a series of maximum optimization problems of a semi-major axis of a circular orbit under the thrust descent fault; adopting an adaptive pseudo-spectral method to solve a circular orbit semi-major axis maximum optimization problem offline, adopting a maximum and minimum method to normalize fault state optimal rescue orbit sample data, normalizing all the data to [-1, 1], adopting an orthogonal least square method to select a radial basis function neural network RBFNN data center, and by selecting a Gaussian basis function as a radial basis function, training a radial basis neural network offline to establish a nonlinear mapping relationship from a fault state to an optimal rescue orbit; and migrating the radial basis function neural network to actual flight of a rocket, and with the fault state of the flight as input, enabling the radial basis function neural network to decide a rescue orbit online.

Description

technical field [0001] The invention relates to the technical field of launch vehicle rescue under failure, in particular to an RBFNN-based rescue orbit decision-making method under rocket thrust drop failure. Background technique [0002] As a rocket power device, the engine is the decisive factor for the flight reliability and safety of the rocket, and its reliability is related to the success or failure of the entire flight mission. In actual flight missions, the failure of the launch vehicle engine and other reasons will easily cause the thrust to drop. If the guidance and control scheme under the nominal ballistic condition is continued, it will be difficult to complete the mission. In order to actively avoid the fall of the payload, it is necessary to re-plan the rescue trajectory and flight trajectory online according to the fault state, so that the payload enters the rescue trajectory. [0003] At present, for the problem of online rescue under thrust drop failure, ...

AI Technical Summary

Problems solved by technology

The above method couples the two problems of rescue orbit decision-making and flight trajectory to optimize together. Since the rescue orbit is unknown, the search space for the optimal solution is large, which affects the calculation efficiency of the online part.

Images