In this paper, the problems of missile formation control in the process of external disturbances, input saturation and unknown obstacles in route planning are researched. Firstly, based on the missile nonlinear dynamic model, the absolute error control model of the missile formation is established by defining the position and speed error auxiliary variable under involving external disturibances. Secondly, on the basis of the nonlinear sliding mode surface based on obstacle avoidance potential function and hyperbolic tangent function, the controller for adaptive anti-saturation and robust missile formation obstacle avoidance is designed by applying adaptive technology and anti-saturation auxiliary system. Finally, Lyapunov stability theory is used to prove that the performance of the designed system is asymptotically stable, and the effectiveness of the designed control strategy is verified by digital simulation.
In order to improve the accuracy and autonomy of aircraft navigation system, a gravity gradient/SINS/starlight integrated navigation method is proposed. In this integrated navigation system, the gravity gradient information is used to correct the position error of SINS, and the starlight information is used to correct the attitude error of SINS to improve the navigation accuracy and autonomy of the aircraft. Regarding overcoming the shortcomings of the traditional cubature Kalman filter algorithm(CKF) which has low accuracy, the random weighted cubature Kalman filter(RWCKF) is used to design the gravity gradient/SINS/starlight integrated navigation system. The simulation results show that the position errors of SINS/starlight integrated navigation system, gravity gradient/SINS integrated navigation system, SAR/SINS/starlight integrated navigation system and the proposed method are 78.1003 m,54.3399 m,39.2776 m and 19.8495 m respectively, which prove that the accuracy of the proposed integrated system is much higher than not only the two subsystems but also the SAR/SINS/starlight integrated navigation system.
In order to realize the compliant docking of the spacecraft and the solar wing, a contact force control strategy based on impedance control is proposed in this paper, and the six-degree-of-freedom parallel attitude adjustment mechanism is token as research purpose. Firstly, the impedance model between the docking member and the member to be docked is established, and the contact force error quantity is converted into the position correction quantity. Secondly, based on the impedance control, an adaptive controller is introduced, which solves the disadvantage of the fixed target parameter of the impedance control and improves the adaptability of the control system to the environment. At the same time, a fuzzy controller is incorporated to adjust the impedance parameter on-line in real time. Finally, a joint simulation is implemented, which verifies the effectiveness of the proposed method by the experimental results based on combining Adams with Simulink.
Aiming at the multi-constraint and strongly time-varying problems of high-speed aircraft during re-entry gliding, this paper combines the online autonomous decision-making advantages of the Deep Deterministic Policy Gradient (DDPG) algorithm to generate avoidance strategies in real time based on threat zone information for dynamic no-fly zone avoidance trajectory planning. In order to further enhance the anti-interference ability of high-speed aircraft to environmental uncertainties, a set of route feature points is selected based on the avoidance trajectory, and a online predictor-corrector guidance method is used to correct the flight status of the high-speed aircraft in real time according to flight mission requirements and terminal constraints, and finally achieve precise guidance of high-speed aircraft. At the same time, in order to verify the effectiveness of the method, this paper carried out corresponding numerical simulation analysis. The results show that the method proposed in this paper can effectively avoid no-fly zones and enhance the adaptability to uncertain factors, and has certain engineering application value.
Aiming at the strong coupling, nonlinearity and severe aerodynamic parameter variations under the complex environment of vehicles, an active disturbance rejection controller design method based on the dragonfly algorithm is proposed. Firstly, a mathematical model based on the three-channel powerless reentry process of vehicles is established. The extended state observer is used to estimate the system state and total disturbance, and then the observed disturbance is compensated. The dragonfly algorithm is employed to optimize the parameters of the active disturbance rejection controller. Finally, validation is shown through a vehicle simulation of six-degree-of-freedom. The results show that the optimized controller parameters have better control accuracy and maintain good control performance even under significant aerodynamic parameter deviations.
A trajectory planning method is proposed for morphing aircraft with threat zones overlapped and full coverage of flight paths, which involves the probability of passing through the threat zone and optimization of variable shape parameters. Based on the idea of hierarchical reinforcement learning, a hierarchical reinforcement learning model for route decision of morphing aircraft is established by configuring the flight environment set, decision options, cost function, Q function and strategies within the options. By training the evaluation network, it can make route decisions for actual scenarios based on the probability of passing through the threat zone. According to the characteristics of the variable shape of the aircraft, the obtained decision results are optimized by parameters to obtain the full process travel trajectory and aircraft shape. The simulation results show that this method can make real-time flight route decisions based on actual situations, and the trajectory and the flight form in overall procedure can be obtained after optimization.
Aiming at the inertial navigation errors rapid increasing in the distributed clusters under the scenario that GNSS signal is unavailable, a collaborative navigation method is proposed in this paper, which just includes total nodes fitted with distributed cluster structure of inertial navigation systems (INSs). A rank-defect constraint Kalman filter is designed to perform collaborative navigation and reduces the errors of the INSs, even if the number of the measurement variables is less than the number of the state variables in the filter. The flight test platform of unmanned air vehicles (UAVs) is established to prove the cooperative navigation system. The results of the flight test show the precision of the position and velocity can be improved by about 2 times when there are none high-precision nodes in the distributed network, and the precision of the position and velocity can be improved by about 10~20 times when there are just one high-precision nodes in the distributed network.
A negotiated differential game optimal evasion strategy for multi pursuit-evasion is proposed for the optimal avoidance in the pursuit-evasion game in capture the flag (CTF). In order to solve this problem, an optimal evasion strategy based on a negotiated differential game is proposed. Firstly, the CTF pursuit-evasion linear model is established and the model reduced in order. Secondly, regarding the cost function based on the energy constraint and the distance between the two sides at the intersection time, Hamilton function is established. Finally, by applying the Hamilton-Jacobi-Isaacs (H-J-I) equations, the optimal evasion strategy based on the negotiated differential game is obtained. The designed strategy is simulated in “two chasing one” and multiplayer pursuit-evasion scenarios. The results show that the flag capture team has the least energy consumption while successfully capturing the flag, which verifies the effectiveness and applicability of the optimal evasion strategy based on a negotiated differential game in this paper.
A multi-channel and end-to-end attitude control method based on deep reinforcement learning is proposed for hypersonic morphing vehicle in the presence of situations by external disturbance and model uncertainty. Firstly, the attitude control model of hypersonic morphing vehicle is established. Secondly, the problem of vehicle attitude control is transformed into a Markov decision process. Furthermore, the training of the agent is implemented, which is based on the twin delayed deep deterministic policy gradient algorithm, and the end-to-end generated flight control instructions are deployed online. Finally, the proposed method's effectiveness and generalizability are confirmed through both basic performance and adaptive simulations.
Regarding the stringent requirements of information confidentiality review in the aerospace field, current manual screening methods are suffering from high costs and insufficient accuracy of keyword matching. An enhanced review framework integrated with large language models is proposed to improve the efficiency and accuracy of confidential information screening. Initially, the characteristics of confidential information are analyzed in the aerospace sector, an architecture that enhances the auditing performance of large language models is introduced in this study, which is combined with dynamic domain-specific expert system prompts to enhance the granularity and accuracy of reviews among multiple perspectives including technical and business confidentiality. By introducing a dynamic system prompt mechanism, the framework is effectively combined the semantic understanding capabilities of large language models with the real-time updating of keywords. Additionally, in order to prevent excessive auditing by the large language model, a hybrid cross-training strategy is developed, which significantly improves the recall rate of confidential information that reaches by 96%. Experiments on a self-developed high quality test set of 1000 entries demonstrates that the proposed method outperforms global open-source large language models by 18% in aerospace classified information inspection tasks.
Inertial measurement unit (IMU) is an important component of anti-aircraft missiles and its state is mainly attained from periodical manual testing, which is not effcient. Aiming at reducing the dependence on periodical testing, time series prediction methods are developed to predict part of the states of IMU by processing existing data. Under the consideration of the small sample size, overlapping segmentation averaging is used for data processing to reduce feature dimension and training difficulty. Long short term memory network (LSTM) is used for subsequent time series prediction. The effect of the proposed model is validated through practical data, which attains a high prediction accuracy while taking little time consumption.
Regarding the performance of less adaptability of the traditional launch vehicle attitude control system which is designed based on the parameters of time retention method that deteriorates drastically when the practical operation condition abnormally deviating from the designed one, a polytope based adaptive control framework is proposed in this paper, which is well suitable for attitude control of rocket’s flight in the aerosphere. The premise of the framework is that the dynamics of rocket in aerosphere can be described by a convex polytope combination through characteristic time operation points retention modelling. On the basis of this assumption, an adaptive control algorithm is designed to solve aerodynamic and rudder efficient uncertainties. Furthermore, parameter-identification based and disturbance-compensation based improvements are proposed, which are capable of dealing with parameter oscillation and model mismatch in normal adaptive control respectively. All three methods are proved to be stable theoretically. These methods can control the attitude of rocket effectively and steadily, and have less parameters to be adjusted, therefore have fairly good industrial prospect.