Research Of Guidance Control And Evaluation For Hypersonic Vehicle

With propose of the United States "Rapid Global Strike" strategic conception,hypersonic aircraft has become a hot research topic in today's military.Due to high flight speed,wide flight range which usually across the airspace,so it will be impacted by the flight environment.Flying missions generally require the aircraft to reach a predetermined location and flight status satisfying various constraints.Due to long flight distance,the aircraft is usually suffers from certain monitoring and interceptions,so these threats must be considered in the guidance system design process.This paper introduces the no-fly zone constraint on the basis of traditional constraints,so that the aircraft can effectively avoid the no-fly zone during flight.In addition,with the deepening of the research,the structure of the hypersonic vehicle is becoming more and more complicated,the precise mathematical model of the aircraft is difficult to obtain,and there are many uncertain factors in the flight process.Therefore,it is necessary to explore the new control method to get rid of the control system design process reliance on accurate mathematical models.With the development of computer technology,the data-based control method has become possible.In this paper,the method research is carried out based on the system data,and the control system design which does not depend on the precise model is realized to a certain extent.The performance evaluation of the guidance and control system is rare in the work of the hypersonic vehicle.In this paper,based on the characteristics of the aircraft,the performance evaluation is carried out on the basis of the guidance control method proposed,so as to complete the guidance control and evaluation.In this paper,the research and development of guidance control and evaluation system are carried out for hypersonic vehicle as the research object,which mainly includes the following aspects:(1)Modeling and characterization for hypersonic vehicle.The guidance system model and the control system model of the hypersonic vehicle are constructed,and the process constraints and the terminal constraints are defined according to the guidance task of hypersonic reentry flight.The characteristics of the hypersonic vehicle were analyzed.During the flight of the hypersonic vehicle,due to the flight speed and the special flight environment,it has far more complex coupling,strong nonlinearity and strong time-varying dynamic characteristics than traditional aircraft.During the flight there is a very strict stance requirement.The flight characteristics of the aircraft need to be analyzed before the data-driven control system is designed.The mechanism of action between each control quantity and state quantity is clarified by the analysis of flight characteristics,and the change relation between each control quantity and state quantity is recorded,which provides reference for the following research work.(2)Guidance method research for hypersonic vehicle considering no-fly zone.The design of the traditional hypersonic vehicle guidance method focuses on various heat flow constraints and terminal state constraints during the flight of the aircraft.The hypersonic vehicle is widely used in the flight process to consider the existence of various interception factors,that is,consider the no-fly zone.This paper analysis the no-fly zone and firstly uses the improved A * algorithm to effectively avoid the no-fly zone and produce a standard avoidance trajectory.In order to introduce process constraints,combined with the idea of optimal control,based on the dynamics of the aircraft and finally guidance conmmand considering no-fly zone can be obtained.(3)Control system design for hypersonic system based on system data.Mainly using neural network control method.The attitude model of the aircraft is analyzed in detail,the unmodeled dynamics in the model are separated,and the unmodeled dynamics of the system are studied.Taking the nonlinear unmodeled dynamics as the research object,a neural network system structure with weights and radial basis functions is constructed.The method of simultaneous perturbation stochastic approximation(SPSA)with variable sliding window is used to estimate the weights.Compared with the traditional method of parameter solving,the SPSA method solves the parameter more quickly.At the same time,the traditional solving methods are difficult to deal with multidimensional parameters.The solving process of SPSA method is not limited by the dimension,and can be conducted more quickly.And the improved SPSA method can further improve the computational efficiency so as to meet the fast changing requirements of hypersonic vehicles.(4)Guidance and control system performance evaluation for hypersonic vehicle.Taking the hypersonic vehicle as the research object,this paper analyzes the performance of the guidance and control system,mainly considering the model confidence evaluation of the aircraft,the establishment of the guidance control performance evaluation system,the scoring strategy for the guidance and control system performance and the adaptive analysis of guidance and control system.(5)Simulation and verification.Simulation and verification are carried out based on the above work.A mathematical model of the aircraft is established so as to build a complete simulation system to verify the effectiveness of guidance and control algorithm.Through the mathematical simulation,this paper presents the rapidity and robustness of the proposed method.The innovations of this paper are:(1)The optimal guidance method for hypersonic vehicle considering no-fly zone is proposed.The no-fly zone constraint during hypersonic flight process is introduced into the guidance instruction process.Finally,the guideline that can meet the evasion requirements of the no-fly zone can be obtained through the theory of optimal control.(2)The design of the control system for hypersonic vehicle was carried out based on the idea of data-driven control.Combining with the characteristics of the aircraft,the nonlinear attitude equation is decomposed and the unmodeled dynamics of the model are dealt by neural network system.The dependence of the control method on the aircraft model is reduced to a certain extent.The existing SPSA method is improved,and the improved method is used to get controller parameters,which improves the efficiency of parameter solving,and the solving process is no longer constrained by the dimensionality of the parameters.(3)Proposed a method to evaluate the performance of hypersonic guidance and control system.Combining with the structure of guidance control system and the characteristic of parameter solving process,the system performance evaluation method with strong pertinence is given,and the guidance and control system is evaluated and verified in all aspects.