Guangyan Xu

Optimal control of UAV elastic formation based on Legendre Pseudospectral Method

The UAV formation reconfiguration problem is described as an optimal control problem in this paper. UAV formation error dynamics are established according to elastic distance vector. The terminal state, control action energy constraints and the constraint of safe distance to avoid colliding are considered during formation reconfiguration. Satisfying all these constraints, the optimal control problem of continuous UAV formation reconfiguration is converted into a nonlinear programming problem. The optimal solution is obtained by Legendre Pseudospectral Method (LPM). The advantage of LPM is to obtain a higher accuracy with fewer nodes. Finally, a three-UAV formation reconfiguration is simulated. It is shown from the results that the ideal formation can be reached within the prescribed time, formation reconfiguration task can be completed and the control standard of minimal formation error and energy consumption can be achieved. The effectiveness and feasibility of the method is verified.

Exact dynamic of satellite relative motion under arbitrary zonal harmonic perturbations

To expand the exact dynamic model of satellite relative motion under J2 perturbation, a new model is established which is exactly up to arbitrary zonal harmonic perturbations. Moreover, a set of differential equations in terms of Reference Satellite Variables (RSV) are obtained to describe the satellite motion around the Earth accurately under arbitrary zonal harmonic perturbations. Since no simplification is applied during the derivation, the result is valid for arbitrary eccentric orbits.

A health diagnosis method for aeroplane structural component based on fault inference engine

As the health status of aeroplane structural components has direct impact on the flight, it is important to diagnose the health status of structural components timely. In this paper, acoustic emission technology is used to monitor the health status of the aeroplane structural component. The acoustic emission health information from the aeroplane structural component is analyzed and disposed. The fault inference engine that bases on rules and the forward chaining control strategies is designed. The K-means clustering analysis algorithm is used to diagnose the aeroplane structural component health status. Experiments show that the method has good performance to diagnoses the status for aeroplane structural components. It presents a good health diagnosis method for aeroplane structural components, which can also be directly applied to other structural systems in machine equipments.

Solving differential game between affine nonlinear system and general nonlinear system

This paper mainly studies differential game problem between an affine nonlinear system and a general nonlinear system, whose control inputs are constrained. In order to solve the problem, a new method is proposed which uses the minimum value principle and the pseudo spectral method. It makes further development and promotion on the basis of semi direct method. Firstly, the system is transformed by the minimum principle. Then the pseudo spectral method is used to solve the whole problem, including the system after transformation by minimum principle and the rest of the general nonlinear system. In order to verify the validity of the method, we use it to solve a pursuit-evasion problem of two nonholonomic systems on the same plane. The simulation results show that this method is feasible to solve the differential game model and has the characteristics of high precision and low computational complexity.

Comparative research on multi-missile cooperative attack between the differential game and proportional navigation method

This paper introduces an approach that involved with multi-missile cooperative attack based on the differential game. By comparing with the traditional proportional navigation method, the differential game shows better performance. Firstly, a mathematical model of multi-missile cooperative attack is established; Secondly, thesis introduces the differential game and proportional navigation method by solving the given problems; Finally, a numerical simulation has been made in Matlab programming environment, which shows that the differential game has more applicability and cooperative tracing ability.

Optimal control of formation reconfiguration for multiple UAVs based on Legendre Pseudospectral Method

The formation reconfiguration problem of the multiple unmanned aerial vehicles (Multi-UAVs) is described as an optimal control problem in this paper. Firstly, the control model of formation flight is constructed with the three-dimensional elastic distance vector, which describes the relationship between the speed and distance of adjacent UAVs in the leader-follower type. Furthermore, the constraints of communication distance and the security anti-collision distance are taken into account, and the formation is modeled to a fuel optimal problem with terminal state restrictions. Then, the optimal control problem of UAV formation flight is transformed into a nonlinear programming problem based on Legendre Pseudospectrum Method. The advantage of LPM is to obtain a higher accuracy with fewer nodes in the process of discretization. Finally, the formation with three UAVs is simulated in Matlab and solved by Tomlab toolkit. The simulation results show that the formation can be well achieved the desired new shape with minor position error and optimal fuel. The effectiveness and feasibility of the method is verified.

Spinning missile attitude estimation method based on the solar vector and MEMS

In this paper, a new method of estimating the attitude to spinning missile was proposed. The author designed the method of using solar vector to correct the gyro bias of MEMS sensor and using the Kalman filter algorithm to estimate the spinning missile attitude. According to the nature of spinning missile, this paper designed a special sun sensor that is used to measure the solar vector replaced the commonly used in MEMS sensor of accelerometer and magnetometer to correct the gyro bias of MEMS sensor. The solar vector measuring device comprises the transparent shell, the photocell array, the real-time clock and the controller. When the sun shines through the transparent shell illuminates the photocell array, the photocell array can convert the solar energy into the electric potential signal. According to the size of the solar energy and the angle of the photocell laying, combining with the accurate time provided by the real time clock, the solar vector of this time is obtained through the main controller processes data. The Kalman filer state equation and observation equation established by the quaternion differential equation is used to estimate the spinning missile attitude. The simulation results show that this method can obtain the higher accuracy attitude angle of spinning missile.

Steady state controller design for aero-engine based on reinforcement learning NNs

An aero-engine optimal steady state controller based on reinforcement learning neural networks (NNs) was proposed in this paper. The presented reinforcement learning NNs can achieve the optimal control objective by constructing two interconnected modules (i.e. action module and critic module). For the state variable models of small perturbation on steady operating points, the double-variable control of an aero-engine is accomplished by two similar backing propagation (BP) NNs. The simulation results show that the presented controller has the perfect performance with the smooth transition process. It not only has strong anti-interference ability and adaptability, but also has excellent robustness to the change of aero-engine model parameters.

Solving nash differential game based on minimum principle and pseudo-spectral method

A novel method based on the minimum principle and the Legendre pseudo-spectral method is proposed to solve the time optimum differential game in this paper. The necessary conditions for the optimal solution of one player are derived according to the minimum principle, which makes the problem of game into the optimal control problem. Then, the Legendre-Lagrange polynomial is employed to interpolate the continuous dynamic equations and the necessary conditions, converting the continuous optimal control problem into the non-linear programming problem solved by the solver snopt. As an illustration, a pursuit-evasion game is formulated with bounded controls and free terminal time. The simulation results show that the method can guarantee the accuracy of the algorithm, and it can greatly improve the operation speed, because it avoids solving the Hamilton-Jacobi equation.

Special inclinations for satellite formations in near circular orbits

The possibility to obtain a natural periodic satellite formation is investigated. This paper shows that such formations under J₂ perturbation have four possible inclinations, which are the inclination at equatorial orbit (i = 0°/180°), inclination at polar orbit (i = 90°), the critical inclination (i = 63.4°/116.6°) and the “succeeding critical inclination” (i = 49.1°/130.9°). Periodic and quasi-periodic relative orbits are introduced based on the special inclinations.