本論文基於模型預測控制（Model Predictive Control, MPC）進行無人搬運車（Automated Guided Vehicle, AGV）之系統開發與設計。在現有之工廠環境中AGV一般都以既定路線上運行來達到定位精確及人車分道增加其工作效率。然而，在以光達導航之AGV控制上是藉由光達進行車輛位置定位以及以當下回授的資訊（車輛位置、車頭角度）來進行車輛下一個狀態之修正（例如：PID控制等）。此一控制方式在整體車輛系統上易產生控制延遲，造成車輛無法準確地跟隨既有之路徑軌跡；為使提高軌跡追蹤之精確度，本文採用模型預測控制，其利用車輛之數學模型來預測狀態、反饋實際軌跡跟蹤誤差及自身角度誤差、優化計算，並得出當前最佳之速度與角速度，且可以在不同的情況採用不同的成本函數，得以讓車輛穩定行駛於軌跡上。
在系統應用上，由於工廠環境不同，可能會有場域過大或是不同樓層之問題，必須要有多張圖資來解決此問題。單一地圖定位導航已經滿足不了現代需求，故本研究提出跨樓層及模型預測控制來進行系統優化。經由軌跡追蹤實驗與跨樓層導航實驗等各項驗證，並採用Pure-pursuit路徑追蹤演算法搭配PID控制器與MPC做比較，驗證其效率提升13.6%以上，證實本論文在實際應用之環境下具有可行性且可信度高。

This study is based on MPC (Model Predictive Control) to proceed the system development and design of AGV (Automated Guided Vehicle). In the existing factory environment, the factory AGV generally follows on the predetermined route to achieve precise positioning and human-vehicle separation to increase efficiency. However, the AGV control based on LiDAR navigation is to proceed the positioning of the vehicle via LiDAR and to proceed adjustment of next state of the vehicle (for example, PID control) according to the current feedback information like vehicle position and yaw angle. PID (Proportional-Integral-Derivative) Control will easily arise control latency on whole vehicle system and cause the AGV to fail to accurately follow the predetermined path. This study applies MPC control in order to increase the precision of trajectory tracking. MPC Control can use the mathematical model of vehicle to predict the state, feedback the error of actual trajectory tracking and the error of yaw angle, optimize calculation, obtain the best velocity and angular velocity under current prediction horizon, and adopt various cost functions under different scenarios in order to make the vehicle drive on the track stably.
In system application, because of the different factory environments, we might have oversized field or multiple levels issues and we need several maps to overcome this situation. Single map navigation is already out of date. Hence, in this research, we submitted cross-field and MPC Control to optimize the system.This framework has been verified through trajectory tracking and cross-floor navigation experiments with 13.6 % efficiency enhancement compared with MPC using Pure-pursuit algorithm and PID Controller. It is confirmed that this paper is feasible and highly credible in the actual application environment.

[1] Mengyuan Chen, and Yue Ren, “MPC based path tracking control for autonomous vehicle with multi-constraints,” Conference on Advanced Mechatronic Systems, pp. 477–482, 2017.
[2] Xiaofeng Liu, Hailin Chen, Chengcheng Wang, Fang Hu and Xianqiang Yang, “MPC Control and Path Planning of Omni-directional Mobile Robot with Potential Field Method,” Conference on Real-time Computing and Robotics（RCAR）, pp. 634-638, 2018.
[3] Matthias Reiter and Dirk Abel, “Two and a Half Carrots - A Versatile and Intuitive Optimisation-Based Path-Following Approach for Road Vehicles*,” Conference on Control and Automation （MED）, pp. 364-370, 2015.
[4] Mohammad Hakim Ahmad Sidi, Khisbullah Hudha, Zulkiffli Abd Kadir and Noor Hafizah Amer, “Modeling and Path Tracking Control of a Tracked Mobile Robot,” IEEE 14th International Colloquium on Signal Processing & its Applications （CSPA 2018）, pp. 72-76, 2018.
[5] Wei Tang, Ming Yang, Chunxiang Wang, Bing Wang, Lunkai Zhang and Fangjie Le,“MPC-Based Path Planning for Lane Changing in Highway Environment,” Conference: 2018 Chinese Automation Congress （CAC）, pp. 1003-1008, 2018.
[6] Chiu-Feng Lin, A.G. Ulsoy and D.J. LeBlanc,“Vehicle dynamics and external disturbance estimation for vehicle path prediction,” IEEE Transactions on Control Systems Technology, vol. 8, issue. 3, pp. 508-518, 2000.
[7] Yiqi Gao, Theresa Lin, Francesco Borrelli, Eric Tseng and Davor Hrovat,“Predictive Control of Autonomous Ground Vehicles With Obstacle Avoidance on Slippery roads,” ASME 2010 Dynamic Systems and Control Conference, pp. 265–272, 2011.
[8] Alexander Liniger, Alexander Domahidi and Manfred Morari,“Optimization-Based Autonomous Racing of 1:43 Scale RC Cars,” Optimal Control Applications and Methods, 36（5）, pp. 628-647, 2015.
[9] Ugo Rosolia, Ashwin Carvalho and Francesco Borrelli,“Autonomous Racing using Learning Model Predictive Control,” Machine Learning （cs.LG）; Optimization and Control （math.OC）, 2017.
[10] Hang Xu and Jin Zhu,“Interval trajectory tracking for AGV based on MPC,” Proceedings of the 38th Chinese Control Conference, pp. 2835-2839, 2019.
[11] 林則宇，「四輪移動車輛結合雲端之路徑追蹤控制」，碩士論文，國立臺灣師範大學，民國105年。
[12] 黃冠智，「基於A*與障礙物區域分析之車輛路徑規劃與避障整合系統」，碩士論文，國立臺灣大學，民國107年。
[13] 徐筱婷，「使用適應性純追蹤基於骨架路徑規劃的多層式修補路徑」，碩士論文，國立臺灣科技大學，民國106年。
[14] 張安宏，「應用感測器融合技術於室內移動機器人同步定位與地圖建構系統開發」，碩士論文，國立臺灣科技大學，民國107年。
[15] 陶忠堅，「無人自行車系統模型和路徑跟隨之控制研究」，博士論文，大葉大學，民國98年。
[16] 陳祐泓，「無人載具路徑跟隨導航之研究」，碩士論文，國立臺北科技大學，民國98年。
[17] 胡凱倫，「具數位地圖能力之室內自動導航車研製」，碩士論文，遠東科技大學，民國103年。
[18] 李益誠，「自動導航貨車之導航與控制系統研究與建置」，碩士論文，國立臺北科技大學，民國102年。
[19] 林晉億，「可自搭電梯之跨樓層機器人」，碩士論文，國立成功大學，民國101年。
[20] Jeong-Gwan Kang, Su-Yong An, and Se-Young Oh,“Navigation Strategy for the Service Robot in the Elevator Environment,” International Conference on Control, Automation and Systems, pp. 1092-1097, 2007.
[21] Reid Simmons et al., “GRACE An Autonomous Robot for the AAAI Robot Challenge,” AIII Mobel Robot Competition, pp. 51-72, 2003.
[22] 袁聖翔，「即時影像辨識之跨樓層物料收發機器人」，碩士論文，國立成功大學，民國101年。
[23] 林文景，「克服樓層障礙之智慧機器人」，碩士論文，國立成功大學，民國95年。
[24] 曾宣橋，「具有人機互動能力之服務型機器人」，碩士論文，國立成功大學，民國98年。
[25] 徐世昌，「郵件機器人之路徑規劃與實現」，碩士論文，國立成功大學，民國96年。
[26] 洪于峻，「可跨樓層遞送物件之模組化載具建造」，碩士論文，國立成功大學，民國106年。
[27] 陳漢忠，「智慧型搭乘電梯機器人」，碩士論文，國立中央大學，民國98年。
[28] Stefan Kohlbrecher, Oskar von Stryk, Johannes Meyer and Uwe Klingauf, “A flexible and scalable SLAM system with full 3D motion estimation,” 2011 IEEE International Symposium on Safety, Security, and Rescue Robotics, pp. 1-6, 2011.
[29] Joseph Redmon, Santosh Divvala, Ross Girshick and Ali Farhadi, “You Only Look Once: Unified, Real-Time Object Detection,” Computer Vision and Pattern Recognition （cs.CV）, 2015.