本文旨在運用基因演算法讓麥克納姆輪移動機器人，以自適應的方式找到每個驅動輪之最佳PID速度控制增益，以改善移動機器人驅動輪之轉速誤差，進而優化移動機器人之軌跡誤差。在參數訓練的過程中融合離線學習與線上學習架構，運用基因演算法搜尋全域最佳解的特性，透過不同控制參數染色體的繁衍，取代原本需要手動調整PID參數之步驟。
在離線學習階段結合自建置之系統模擬器與基因演算法，以模擬的方式進行參數訓練，有效的縮短訓練時間，並且將最佳參數聚焦於可能之候選區間。而在線上學習之訓練結構中，則以粗微調兩階段方式進行訓練，並且分別利用實際移動機器人之驅動輪轉速誤差以及移動機器人之移動軌跡誤差，做為基因演算法之損失函數分數，反饋基因序列之控制成效，讓移動機器人於運動過程中找到每一個驅動輪之最佳PID速度控制增益，達成自主學習之PID控制器設計。
本移動機器人之軌跡控制，是在程式控制端設計循環式狀態機，搭配麥克納姆輪移動機器人之反運動學公式，以梯形加減速進行路徑規劃，最後依照軌跡之控制點資訊依序傳送控制指令，實現麥克納姆輪移動機器人之軌跡控制。

This thesis aims to find the best PID speed control gains for each driving wheel by using genetic algorithms so as to improve the trajectory control error of a Mecanum wheeled mobile robot. The process of PID parameters tuning is done in two stages, an offline learning and followed by an online learning. Genetic algorithms are applied to search for the best PID parameters without manual adjustment.
In the offline learning stage, a simulator system is built and the genetic algorithm is used to conduct parameter search. In this way, it effectively shortens the training time and focuses the best parameters on the possible candidate interval. In the online learning stage, the training process is consistent of coarse and fine parameters adjustment. By using the driving wheel speed error of the actual mobile robot as the loss function in using the genetic algorithm, the best PID speed controller of mobile robot can be found during the online training.
The trajectory control of the Mecanum wheeled mobile robot is performed by using a sequential state machine and the inverse differential kinematics. The planned path is followed by trapezoidal acceleration and deceleration profile, and the control points of the trajectory are tracking in a sequence to realize the trajectory control.

[1] J. Park, S. Kim, J. Kim and S. Kim, "Driving control of mobile robot with Mecanum wheel using fuzzy inference system," ICCAS 2010, Gyeonggi-do, Korea (South), 2010, pp. 2519-2523, doi: 10.1109/ICCAS.2010.5670241.
[2] X. Lu, X. Zhang, G. Zhang and S. Jia, "Design of adaptive aliding mode controller for four-Mecanum wheel mobile robot," 2018 37th Chinese Control Conference (CCC), Wuhan, China, 2018, pp. 3983-3987, doi: 10.23919/ChiCC.2018.8483388.
[3] J. S. Keek, S. L. Loh and S. H. Chong, "Comprehensive development and control of a path-trackable Mecanum-wheeled robot," in IEEE Access, vol. 7, pp. 18368-18381, 2019, doi: 10.1109/ACCESS.2019.2897013.
[4] Fahmizal and C. Kuo, "Trajectory and heading tracking of a mecanum wheeled robot using fuzzy logic control," 2016 International Conference on Instrumentation, Control and Automation (ICA), Bandung, Indonesia, 2016, pp. 54-59, doi: 10.1109/ICA.2016.7811475.
[5] G. Lin and G. Liu, "Tuning PID controller using adaptive genetic algorithms," 2010 5th International Conference on Computer Science & Education, Hefei, China, 2010, pp. 519-523, doi: 10.1109/ICCSE.2010.5593559.
[6] M. Korkmaz, Ö. Aydoğdu and H. Doğan, "Design and performance comparison of variable parameter nonlinear PID controller and genetic algorithm based PID controller," 2012 International Symposium on Innovations in Intelligent Systems and Applications, Trabzon, Turkey, 2012, pp. 1-5, doi: 10.1109/INISTA.2012.6246935.
[7] R. Wen and M. Tong, "Mecanum wheels with Astar algorithm and fuzzy PID algorithm based on genetic algorithm," 2017 International Conference on Robotics and Automation Sciences (ICRAS), Hong Kong, China, 2017, pp. 114-118, doi: 10.1109/ICRAS.2017.8071927.
[8] C. Vlachos, J. T. Evans and D. Williams, "PI controller tuning for multivariable processes using genetic algorithms," Second International Conference On Genetic Algorithms In Engineering Systems: Innovations And Applications, 1997, pp. 43-49, doi: 10.1049/cp:19971153.
[9] X. Meng and B. Song, "Fast genetic algorithms used for PID parameter optimization," 2007 IEEE International Conference on Automation and Logistics, 2007, pp. 2144-2148, doi: 10.1109/ICAL.2007.4338930.
[10] J. Juang, M. Huang and W. Liu, "PID control using presearched genetic algorithms for a MIMO system," in IEEE Transactions on Systems, Man, and Cybernetics, Part C (Applications and Reviews), vol. 38, no. 5, pp. 716-727, Sept. 2008, doi: 10.1109/TSMCC.2008.923890.
[11] R. A. Krohling and J. P. Rey, "Design of optimal disturbance rejection PID controllers using genetic algorithms," in IEEE Transactions on Evolutionary Computation, vol. 5, no. 1, pp. 78-82, Feb 2001, doi: 10.1109/4235.910467.
[12] J. Han, "From PID to active disturbance rejection control," in IEEE Transactions on Industrial Electronics, vol. 56, no. 3, pp. 900-906, March 2009, doi: 10.1109/TIE.2008.2011621.
[13] Kiam Heong Ang, G. Chong and Yun Li, "PID control system analysis, design, and technology," in IEEE Transactions on Control Systems Technology, vol. 13, no. 4, pp. 559-576, July 2005, doi: 10.1109/TCST.2005.847331.
[14] J. J. Grefenstette, "Optimization of control parameters for genetic algorithms," in IEEE Transactions on Systems, Man, and Cybernetics, vol. 16, no. 1, pp. 122-128, Jan. 1986, doi: 10.1109/TSMC.1986.289288.
[15] M. Srinivas and L. M. Patnaik, "Adaptive probabilities of crossover and mutation in genetic algorithms," in IEEE Transactions on Systems, Man, and Cybernetics, vol. 24, no. 4, pp. 656-667, April 1994, doi: 10.1109/21.286385.
[16] G. Rudolph, "Convergence analysis of canonical genetic algorithms," in IEEE Transactions on Neural Networks, vol. 5, no. 1, pp. 96-101, Jan. 1994, doi: 10.1109/72.265964.
[17] 施慶隆、李文猶,機電整合與控制-多軸運動設計與應用,第三版,全華書局股份有限公司,2015。