Parameter Optimization Of Tracking Controller For Soft Landing Of Lunar Probe

An important way to obtain scientific data in deep space exploration is to land the target celestial body and return to earth after sampling by the lander.In recent years,with the development of science and technology,the moon,as the nearest celestial body to the earth,has been gradually found in many valuable exploration areas.The exploration of the moon needs strong technical support,and soft landing technology is the most important.In this paper,the soft landing trajectory tracking of lunar probe is studied in detail.The main work is as follows.On the basis of analyzing the soft landing process of the lunar probe,the orbit constraints of the landing phase are established by synthesizing the information of each landing stage.Further,in the polar coordinate system,the two body model of lunar soft landing is established.Based on the research of pseudo-spectral method,the discrete model of soft landing trajectory optimization is established by using Gaussian pseudospectral method.The optimal control problem of continuous system is transformed into a finite-dimensional parametric optimization problem.The optimal nominal trajectory is obtained by simulation and analysis.Aiming at the problem of tracking the optimal nominal trajectory of lunar probe soft landing,a basic LQR controller is first designed by introducing a linear quadratic regulator to realize trajectory tracking.Aiming at the problem of selecting Q matrix of the basic LQR controller,the selection of Q matrix is further optimized on the basis of the reinforcement learning AC algorithm.Finally,by comparing the simulation results of the basic LQR controller with the optimal linear controller,the effectiveness of the proposed algorithm is verified,and the tracking accuracy of the optimal linear controller is greatly improved.In order to solve the problem that network initialization has a great impact on the effect of Q matrix selection in AC algorithm,this paper continues to improve the traditional particle swarm optimization algorithm from the perspective of bionics,and proposes an improved particle swarm optimization algorithm which combines fitness factor and improvement factor as a new inertial factor for the optimization of Q-matrix of track tracking controller.The improved particle swarm optimization algorithm is superior to the traditional particle swarm optimization algorithm in terms of stability and adaptability.Finally,through Compared with the simulation results of traditional particle swarm optimization algorithm and improved particle swarm optimization algorithm,the effectiveness of the proposed algorithm is verified.