PID Intelligent Control Based On ACPSO-WFLN For Control And Optimization Of Quadrotor

Quad-rotor aircraft has a large possibility of unstable flight because it is susceptible to the external environment.Therefore,it is of great practical significance to seek excellent control algorithms to ensure the stability of the aircraft pose.In this thesis,a PID intelligent control algorithm based on ACPSO and WFLN is proposed to solve the problem that PID control algorithms require manually adjusting control parameters and optimal control is difficult to achieve.And from the control system analysis and aircraft pose control of the two aspects of the analysis,the main research contents of the thesis is as follows:1.An improved WFLN network model is proposed.Aiming at the limitation of generalization performance of wavelet neural network,this thesis improves the structure of WNN network by adding a direct connection between input and output vectors,and proposes an improved wavelet functional link neural network with good parallel processing capability.Firstly,WFLN network model is constructed by adding a linear connection unit between the input layer and the output layer of the wavelet neural network.Secondly,the wavelet basis function is used as the transfer function of the hidden layer of WFLN,and thegradient correction method is conducted to train the parameters of the network model.Finally,the results show that this improved network can reduce the number of neurons needed in the hidden layer compared with WNN network,and effectively improve the overall parallel computing ability and convergence speed of the network.2.ACPSO algorithm optimizes WFLN network model.In view of the deficiency of WFLN network in gradient training process that initial weights cannot be determined and it is easy to fall into local extremum,this thesis proposes an adaptive chaotic particle swarm optimization algorithm with efficient optimization ability to optimize initial parameters of WFLN network.Four standard test functions were used to compare and analyze the optimization capability of traditional GA algorithm,PSO algorithm and ACPSO algorithm proposed in this thesis.The results show that ACPSO not only has less iteration steps than PSO algorithm,but also has more efficient dynamic balance advantages.Then,ACPSO algorithm is used to optimize the initial connection weight of WFLN network,and the ACPSO-WFLN network algorithm model is constructed.Simulation results show that the ACPSO-WFLN network model has relatively low root mean square error,and has a good overall optimization ability.3.PID intelligent control based on ACPSO-WFLN is analyzed.Considering that traditional PID control,single neuronal PID Controlle,traditional multilayer feedforward Neuronal PID Controlle have slowconvergence speed and high overshoot,ACPSO-WFLN network model proposed in this thesis is applied to traditional PID algorithm to improve the PID parameter tuning efficiency.Then,the three control algorithms are compared with algorithm proposed in this thesis by comparative analysis of control system simulation.The simulation results show that the proposed combination optimization model has the advantages of fast convergence and small overshoot in the control field,providing a more effective control strategy for the four-rotor attitude control.4.Research on intelligent control of quadrotor based on ACPSO-WFLN-PID.According to the working principle and modeling analysis of the quadrotor,the ACPSO-WFLN-PID intelligent algorithm proposed in this thesis is applied to the field of quadrotor control.Firstly,a position and attitude intelligent controller based on ACPSO-WFLN-PIDNN is established.Then,simulation experiments are carried out from three aspects: step response of aircraft,anti-jamming performance and track tracking performance.The results show that compared with the traditional PID control algorithm,the accuracy of the control algorithm is improved by 58.23%.Compared with the WFLN-PID control algorithm,the accuracy of the control algorithm is improved by 22.19%,which verifies the good self-adaptability and effectiveness of the intelligent control strategy proposed by this thesis.