A method aiming at the formation control problem during the reentry of multi-unpowered vehicles is discussed in this paper. Based on second-order system consistency theory of virtual leader, a distributed formation keeping and control algorithm switching of adjustable assembly time is proposed by selecting optimal coordination variables and introducing a special transition link. The altitude of multi-vehicles are allowed to be changed in a certain range by using this proposed solution that considers the flight ability under the restraint of multi-vehicles formation mission and then realizes the accurate online control of the relative transverse and longitudinal range. Simulation results show that the proposed algorithm can achieve the desired control effect in scenarios of formation assembly and keeping, formation configuration switching and formation lateral maneuver. The algorithm takes into account the effects of the earth’s curvature and rotation on the reentry dynamics, and can adapt to random aerodynamic coefficients and atmosphere density pulling by good robustness.
A reliable target trajectory tracking controller is designed to address the issues of external wind disturbances and model uncertainties encountered during flight of quadrotor unmanned aerial vehicle (UAV) within natural environments. An adaptive sliding mode control method based on recurrent neural network (RNN) is proposed. Regarding starting from the dynamics model of quadrotor UAV, the fully actuated and underactuated subsystems are considered separately. By combining external wind disturbances with model uncertainties and then integrating into total disturbance terms, a recurrent neural network is used to estimate the total disturbance terms adaptively. Based on Lyapunov theory, a feedback compensation adaptive sliding mode controller is designed. Thus, the effectiveness of the proposed control method is fully verified through theoretical and comparative simulations.
Based on the mid-flight characteristics of ballistic missiles, a controllable coordinated interception terminal guidance frame for attack time based on hit point prediction is proposed. By adopting the model of predicting hit points, high-speed targets can be effectively transformed into fixed or low-speed targets. By considering the real-time requirements, a midpoint prediction algorithm based on the combination of simulated annealing and binary method is proposed. The hit point prediction is fairly accurate, the calculation speed is fast, the computational load is small, and the reduction of computational burden of missile borne computers is achieved in effectiveness. A "leader follower" structure is applied to solution, the lead missile adopts proportional guidance to quickly approach the predicted midpoint and the remaining flight time of the follower missile is changed by adjusting its own flight trajectory, which makes remaining flight time of follower missile equal to the estimated remaining flight time of the lead missile. The simulation results show that three interceptors in this guidance frame are enabled to hit the target simultaneously.
An online trajectory optimization and guidance algorithm based on convex optimization is proposed for the soft landing maneuver guidance problem under engineering constraints such as thrust pointing constraint, speed direction constraint and thrust change rate constraint. Based on the optimal control theory, the influence of thrust pointing constraint on the relaxation convexization of non-convex constraint is analyzed, and the Lagrange interpolation method is used to eliminate the endogenous fluctuation of the guidance command. The multi-state multi-flight test is implemented, which is based on the reusable flight test platform. The flight test results show that the algorithm can adapt to the planning and guidance requirements under various constraints and conditions and has the potential to be further served as engineering practice.
A continuous active disturbance rejection sliding mode control strategy based on finite-time convergent extended state observer technology is proposed for solving the active disturbance rejection fixed-point control problem of quadrotor unmanned aerial vehicle (UAV) during ultra-high voltage electrical testing. By firstly considering that the quadrotor UAV is difficult to accurately obtain variable information such as linear velocity and angular velocity during flight and are susceptible to external time-varying disturbances, an extended state observer with finite-time convergence is used to estimate unknown linear velocity, angular velocity and lumped disturbance terms. Moreover, the proposed strategy can effectively reduce the position and attitude tracking errors, eliminate the chattering phenomenon of sliding mode control and achieve high-precision fixed-point control of quadrotor UAV. Finally, the effectiveness of the proposed control strategy is fully verified through theory and simulation.
Aiming at the problem that magnetometer or sun sensors can only estimate the attitude of two axes, the observed quantity is presented by using the method of taking advantage of the observation vector measured by sun sensors and magnetometer, the attitude error is gained by using sun vector which is compared with the geomagnetic field vector estimated based on the orbit and the attitude angular velocity of the satellite. The integrated attitude of the gyro is corrected by the attitude error estimated based on the sun vector and the geomagnetic field vector, and the gyro drift is estimated based on PI filter. The simulation result shows that the attitude determination precision reaches about 1°.
In response to the complex and uncertain conditions faced by solar sail spacecraft during in-orbit flight, a deep reinforcement learning-based integrated algorithm for trajectory design and robust guidance is proposed. The uncertainties conditions acted as solar radiation pressure model uncertainty, navigation error, control execution error and randomly triggered safety events are incorporated into the Markov decision process modeling of solar sail spacecraft in-orbit flight, based on the algorithm proposed regarding orbital dynamics of solar sail spacecraft. A reward function reflecting the optimization of solar sail energy supply is designed by the minimum solar phase angle, and training is conducted by using the proximal policy optimization algorithm to achieve the optimization design of solar sail spacecraft trajectories and robust guidance under complex and uncertain conditions. This algorithm is applied to the heliocentric transfer mission of a solar sail spacecraft exploring the near-Earth asteroid 2019 GF1. Simulation results show that the terminal arrival precision of nominal trajectory tracking flight under uncertain conditions can be decreased and the solar phase angle along the trajectory is reduced by using this new algorithm.
A hybrid control method combining feed-forward compensation with closed-loop feedback based on the target relative motion trajectory is proposed in this paper, which is aiming at significantly improving the satellite dynamic pointing tracking control precision. Firstly, the dynamic error coefficient method is adopted to analyze the attitude tracking errors, and power transfer function is established to analyze attitude angular velocity errors caused by sensor noise. Next, a feed-forward compensation control strategy calculating the desired line-of-sight angular velocity and angular acceleration by estimation of the target relative motion is proposed, which resolves the challenge of space-borne detection equipment that can only measure target pointing deviations and doesn’t directly perform hybrid control. Finally, the system stability and attitude tracking performance are validated through mathematical simulations, and ground-based air-bearing experiments are implemented to test the controller's tracking performance. This attitude controller reaches high-precision during dynamic tracking by different input commands without altering the closed-loop feedback bandwidth.
Aiming at the demand of high-precision detection for the distributed spatial optical interference, a cooperative control system design scheme is proposed in this paper to address the problem of insurmountable internal and external interferences in the cooperative pointing of distributed detector platforms. Firstly, dynamics modeling of the distributed detectors and analysis of the interference mechanism are implemented. Then, on the basis of that, a distributed interference observer and an RBF network interference learning observer are designed for analyzing and separating the perturbation variables from other state quantities of the system, which are used to reduce the influence of perturbation uncertainties on the system stability. Furthermore, an attitude tracking controller is designed to eliminate uncertain perturbations of the system. Finally, simulation results show the effectiveness of the proposed design scheme and inspire the design of cooperative control system for distributed detector platforms.
In order to achieve high-speed data transmission between the user satellite antennas and relay satellite, single channel monopulse tracking systems are used to realize high-precision tracking. The design of auto tracking system based on integration of Ka-band tracking receiver and Servo controller is described in this paper, where TE21 mode coupler is used on Ka-band antenna. Tracking receiver model is established on the nonlinear dynamics model. Auto-tracking pointing error is provided by the model. Comparing the simulation with experiment results, the math model is correct and reliable. The Auto-tracking system is test on far-field. Tracking receiver’s S curve, dynamic performance and pointing error are applied to test. The experiment results demonstrate the auto-tracking system function and performance meets the requirements of task by precision pointing.
In order to improve the reliability of the liquid rocket engine, aiming at solving the liquid rocket engine fault diagnosis problem, a kind of fault diagnosis model based on the kernel principal component analysis (KPCA) and the ICOA-BP algorithm is proposed in this paper. The feature extraction and dimensionality degradation of measured parameters are implemented by using KPCA algorithm which ensures sufficient features amount extracted and coming together by reducing the complexities of the data and the computational cost. And an improved Coati optimization algorithm (ICOA) is proposed to optimize the BP neural network, aiming at improving the diagnostic accuracy of the BP neural network. The algorithm is validated by using liquid-oxygen-methane rocket engine test data, and the experimental results show that the ICOA-BP algorithm exhibits faster convergence speed and higher optimization finding accuracy compared to the COA-BP algorithm. The diagnostic accuracy of ICOA-BP algorithm can reach 96.5% on the data extracted from KPCA features, which is respectively 3.5% and 3% higher than the diagnostic accuracy of BP neural network and Support Vector Machine (SVM). Compared with particle swarm algorithm (PSO) and genetic algorithm (GA), ICOA-BP algorithm demonstrates better searching ability for the global optimal solution.
