新型移動平衡機器人之運動控制設計與實作

本論文設計出一台移動平衡機器人，機器人的移動方式傳承麥克納姆輪的特性可以具備全方位移動的能力，突破一般常見的移動平衡機器人無法側向移動的缺陷，使得所提出之機器人能夠穿越狹窄的空間。在硬體實現方面，本論文選用直流無刷馬達提供機器人移動的動力，使用MPU6050姿態感測器來取得機器人的姿態信息，與使用NVIDIA Jetson Nano來當作機器人的運算處理核心。在軟體實現方面，本論文基於機器人操作系統ROS接收馬達之角度資料與傳送馬達命令，並且利用Ros bag功能記錄實驗數據，將數據圖形化以利觀察整理，同時提出了兩種不同的控制器來進行驗證，分別為PID運動控制器和模糊滑動模式運動控制器。經由實驗結果實際驗證與比較，所提出之兩種運動控制器均可以成功地控制所設計之移動平衡機器人，且模糊滑動模式運動控制器可以獲得較佳之控制結果。進一步，提出了遠端遙控與軌跡追蹤操控兩種方式，讓使用者透過PS2搖桿來即時控制機器人，或是讓機器人會依照預設的軌跡來移動，也都獲得不錯的控制結果。

關鍵字

移動平衡機器人；全方位移動；機器人操作系統；運動控制器

並列摘要

This thesis aims to design and implement a self-balancing mobile robot. The moving method of the proposed robot inherits the characteristics of the Mecanum wheel, so that it can have omni-directional movement capabilities. In the hardware implementation, four BLDC motors are equipped to provide the power of robot movement, MPU6050 attitude sensor is used to measure the robot attitude information, and the NVIDIA Jetson Nano is used as the robot's computing core. In the software implementation, this thesis receives the motors encoder data and sends the motor control commands based on the Robot Operating System (ROS). Meanwhile, the Ros bag function is setup to record the experimental data, and graph the data to observation and sorting. For the controller design, a PID control system and a fuzzy sliding-mode control system are proposed for the self-balancing mobile robot. The experimental results show that both controllers are able to control the robot successfully and the fuzzy sliding-mode controller has better control results. Further, this thesis proposes the remote control and the trajectory tracking control, so that the user can manipulate the robot through the PS2 joystick or the robot can follow a trajectory to move. The experimental results show that both of remote control and trajectory tracking control can achieve favorable control results and keep the robot in balance.