# American Institute of Mathematical Sciences

October  2017, 13(4): 2067-2091. doi: 10.3934/jimo.2017032

## Comparison of GA and PSO approaches for the direct and LQR tuning of a multirotor PD controller

 Faculty of Engineering and Architecture, Cittadella Universitaria -Enna, Italy

* Corresponding author

Received  September 2015 Revised  October 2016 Published  April 2017

In the present paper different approaches for flight control stabilization tuning of a prototype of Unmanned Aerial Vehicle (UAV) are investigated. The mathematical model of a multirotor airframe is presented and its dynamical system is described by means of Newton-Euler equations for a rigid body. The proposed controller systems for the stabilization of the altitude and the attitude of the multirotor around its hovering configuration are based on Proportional Derivative (PD) regulator tuned by optimization techniques such as Particle Swarm Optimization (PSO) and Genetic Algorithm (GA) or, for the linearized system, by means of Linear Quadratic Regulator (LQR). In this case, PSO and GA are applied to set entries of LQR scheme. Control flight simulation results have been compared in terms of minimization of work carried out by the control actions. Performed simulations pointed out that all approaches lead to zero the error of the position along gravity acceleration direction, stop the rotation of UAV around body axes and stabilize the multirotor. The direct PSO tuning of the PD parameters (PD-PSO) shows better results in terms of errors, performance and speed of convergence.

Citation: Valeria Artale, Cristina L. R. Milazzo, Calogero Orlando, Angela Ricciardello. Comparison of GA and PSO approaches for the direct and LQR tuning of a multirotor PD controller. Journal of Industrial and Management Optimization, 2017, 13 (4) : 2067-2091. doi: 10.3934/jimo.2017032
##### References:

show all references

##### References:
Convergence trend for GA-PD algorithm: 10 runs are represented
Convergence trend for PSO-PD algorithm: 10 runs are represented
Convergence trend for GA-LQR-PD algorithm: 10 runs are represented
Convergence trend for PSO-LQR-PD algorithm: 10 runs are represented
Root locus of the controlled systems
Comparison between altitude trend versus time in GAPD case (red line) and in PSO-PD case (black line)
Comparison between roll trend versus time in GA-PD case (red line) and in PSO-PD case (black line)
Comparison between pitch trend versus time in GA-PD case (red line) and in PSO-PD case (black line)
Comparison between yaw trend versus time in GA-PD case (red line) and in PSO-PD case (black line)
Comparison between thrust trend versus time in GAPD case (red line) and in PSO-PD case (black line)
Comparison between roll moment trend versus time in GA-PD case (red line) and in PSO-PD case (black line)
Comparison between pitch moment trend versus time in GA-PD case (red line) and in PSO-PD case (black line)
Comparison between yaw moment trend versus time in GA-PD case (red line) and in PSO-PD case (black line)
Comparison between altitude trend versus time in GA-LQR-PD case (red line) and in PSO-LQR-PD case (black line)
Comparison between roll trend versus time in GALQR-PD case (red line) and in PSO-LQR-PD case (black line)
Comparison between pitch trend versus time in GALQR-PD case (red line) and in PSO-LQR-PD case (black line)
Comparison between yaw trend versus time in GALQR-PD case (red line) and in PSO-LQR-PD case (black line)
Comparison between thrust trend versus time in GALQR-PD case (red line) and in PSO-LQR-PD case (black line)
Comparison between roll moment trend versus time in GA-LQR-PD case (red line) and in PSO-LQR-PD case (black line)
Comparison between pitch moment trend versus time in GA-LQR-PD case (red line) and in PSO-LQR-PD case (black line)
Comparison between yaw moment trend versus time in GA-LQR-PD case (red line) and in PSO-LQR-PD case (black line)
Air turbulence: altitude time history
Air turbulence: roll time history
Air turbulence: pitch time history
Air turbulence: yaw time history
Air turbulence: thrust time history
Air turbulence: roll moment time history
Air turbulence: pitch moment time history
Air turbulence: yaw moment time history
Proportional and Derivative coefficients gained by GA-PD and PSO-PD algorithms
 GA-PD PSO-PD i $K_{i, P}$ $K_{i, D}$ $K_{i, P}$ $K_{i, D}$ $z$ $38327$ $923.59$ $46879$ $1000$ $\phi$ $19776$ $632.13$ $20000$ $1000$ $\theta$ $16918$ $246.83$ $20000$ $1000$ $\psi$ $18142$ $162.16$ $20000$ $1000$
 GA-PD PSO-PD i $K_{i, P}$ $K_{i, D}$ $K_{i, P}$ $K_{i, D}$ $z$ $38327$ $923.59$ $46879$ $1000$ $\phi$ $19776$ $632.13$ $20000$ $1000$ $\theta$ $16918$ $246.83$ $20000$ $1000$ $\psi$ $18142$ $162.16$ $20000$ $1000$
Proportional and Derivative coefficients gained by GA-LQR-PD and PSO-LQR-PD algorithm
 GA-LQR-PD PSO-LQR-PD i $K_{i, P}$ $K_{i, D}$ $K_{i, P}$ $K_{i, D}$ $z$ $32.89$ $33.013$ $707.11$ $207.48$ $\phi$ $447.21$ $35.386$ $447.21$ $42.964$ $\theta$ $252.13$ $32.336$ $447.21$ $42.964$ $\psi$ $310.79$ $41.579$ $447.21$ $50.254$
 GA-LQR-PD PSO-LQR-PD i $K_{i, P}$ $K_{i, D}$ $K_{i, P}$ $K_{i, D}$ $z$ $32.89$ $33.013$ $707.11$ $207.48$ $\phi$ $447.21$ $35.386$ $447.21$ $42.964$ $\theta$ $252.13$ $32.336$ $447.21$ $42.964$ $\psi$ $310.79$ $41.579$ $447.21$ $50.254$
Comparison of objective function values at the final iteration step
 GA-PD PSO-PD GA-LQR-PD PSO-LQR-PD min 1.5442 1.5152 51.846 30.978 max 1.6271 1.5159 957.32 49.788 mean 1.5799 1.5156 457.47 37.234 std 0.02871 0.00036 348.55 7.3161 ISE 0.00358 0.00356 0.05474 0.00950 IAE 0.04889 0.05031 0.33813 0.13091
 GA-PD PSO-PD GA-LQR-PD PSO-LQR-PD min 1.5442 1.5152 51.846 30.978 max 1.6271 1.5159 957.32 49.788 mean 1.5799 1.5156 457.47 37.234 std 0.02871 0.00036 348.55 7.3161 ISE 0.00358 0.00356 0.05474 0.00950 IAE 0.04889 0.05031 0.33813 0.13091
 [1] Hong Niu, Zhijiang Feng, Qijin Xiao, Yajun Zhang. A PID control method based on optimal control strategy. Numerical Algebra, Control and Optimization, 2021, 11 (1) : 117-126. doi: 10.3934/naco.2020019 [2] Bin Li, Kok Lay Teo, Cheng-Chew Lim, Guang Ren Duan. An optimal PID controller design for nonlinear constrained optimal control problems. Discrete and Continuous Dynamical Systems - B, 2011, 16 (4) : 1101-1117. doi: 10.3934/dcdsb.2011.16.1101 [3] Bin Li, Xiaolong Guo, Xiaodong Zeng, Songyi Dian, Minhua Guo. An optimal pid tuning method for a single-link manipulator based on the control parametrization technique. Discrete and Continuous Dynamical Systems - S, 2020, 13 (6) : 1813-1823. doi: 10.3934/dcdss.2020107 [4] Simone Fiori, Italo Cervigni, Mattia Ippoliti, Claudio Menotta. Synchronization of dynamical systems on Riemannian manifolds by an extended PID-type control theory: Numerical evaluation. Discrete and Continuous Dynamical Systems - B, 2022  doi: 10.3934/dcdsb.2022047 [5] H. W. J. Lee, Y. C. E. Lee, Kar Hung Wong. Differential equation approximation and enhancing control method for finding the PID gain of a quarter-car suspension model with state-dependent ODE. Journal of Industrial and Management Optimization, 2020, 16 (5) : 2305-2330. doi: 10.3934/jimo.2019055 [6] Andrew D. Lewis, David R. Tyner. Geometric Jacobian linearization and LQR theory. Journal of Geometric Mechanics, 2010, 2 (4) : 397-440. doi: 10.3934/jgm.2010.2.397 [7] Petra Csomós, Hermann Mena. Fourier-splitting method for solving hyperbolic LQR problems. Numerical Algebra, Control and Optimization, 2018, 8 (1) : 17-46. doi: 10.3934/naco.2018002 [8] Ramesh Devarapalli, Biplab Bhattacharyya. A novel hybrid AGWO-PSO algorithm in mitigation of power network oscillations with STATCOM. Numerical Algebra, Control and Optimization, 2021, 11 (4) : 579-611. doi: 10.3934/naco.2020057 [9] Marc Puche, Timo Reis, Felix L. Schwenninger. Funnel control for boundary control systems. Evolution Equations and Control Theory, 2021, 10 (3) : 519-544. doi: 10.3934/eect.2020079 [10] Qun Lin, Ryan Loxton, Kok Lay Teo. The control parameterization method for nonlinear optimal control: A survey. Journal of Industrial and Management Optimization, 2014, 10 (1) : 275-309. doi: 10.3934/jimo.2014.10.275 [11] Alain Bensoussan, John Liu, Jiguang Yuan. Singular control and impulse control: A common approach. Discrete and Continuous Dynamical Systems - B, 2010, 13 (1) : 27-57. doi: 10.3934/dcdsb.2010.13.27 [12] Cristiana J. Silva, Helmut Maurer, Delfim F. M. Torres. Optimal control of a Tuberculosis model with state and control delays. Mathematical Biosciences & Engineering, 2017, 14 (1) : 321-337. doi: 10.3934/mbe.2017021 [13] Victor Ayala, Adriano Da Silva, Luiz A. B. San Martin. Control systems on flag manifolds and their chain control sets. Discrete and Continuous Dynamical Systems, 2017, 37 (5) : 2301-2313. doi: 10.3934/dcds.2017101 [14] Changjun Yu, Shuxuan Su, Yanqin Bai. On the optimal control problems with characteristic time control constraints. Journal of Industrial and Management Optimization, 2022, 18 (2) : 1305-1320. doi: 10.3934/jimo.2021021 [15] Yves Dumont, Frederic Chiroleu. Vector control for the Chikungunya disease. Mathematical Biosciences & Engineering, 2010, 7 (2) : 313-345. doi: 10.3934/mbe.2010.7.313 [16] Andrew D. Lewis. Linearisation of tautological control systems. Journal of Geometric Mechanics, 2016, 8 (1) : 99-138. doi: 10.3934/jgm.2016.8.99 [17] Dušan M. Stipanović, Christopher Valicka, Claire J. Tomlin, Thomas R. Bewley. Safe and reliable coverage control. Numerical Algebra, Control and Optimization, 2013, 3 (1) : 31-48. doi: 10.3934/naco.2013.3.31 [18] Thalya Burden, Jon Ernstberger, K. Renee Fister. Optimal control applied to immunotherapy. Discrete and Continuous Dynamical Systems - B, 2004, 4 (1) : 135-146. doi: 10.3934/dcdsb.2004.4.135 [19] Yuri B. Gaididei, Carlos Gorria, Rainer Berkemer, Peter L. Christiansen, Atsushi Kawamoto, Mads P. Sørensen, Jens Starke. Stochastic control of traffic patterns. Networks and Heterogeneous Media, 2013, 8 (1) : 261-273. doi: 10.3934/nhm.2013.8.261 [20] Ellina Grigorieva, Evgenii Khailov. Optimal control of pollution stock. Conference Publications, 2011, 2011 (Special) : 578-588. doi: 10.3934/proc.2011.2011.578

2020 Impact Factor: 1.801