本研究利用模糊類神經網路和卡爾曼濾波器來控制六足機器人，進行沿牆避障追蹤並且有效率的調整姿態使得機器人在運動過程中有很高的穩定性。根據架設在機器人側邊的超音波感測器所得到距離值，經由換算可以得知機器人與牆壁的相對角度位置，而模糊類神經網路的輸入為角度位置，輸出為機器人兩側足部的擺動幅度，經由兩邊足部擺動的差異可以調整機器人的前進方向，使得機器人可以在複雜的環境中避開障礙物。除了避障之外，保持機器人身體的穩定度也是很重要的議題，但使用加速度計所測量的姿態傾角，會因為響應靈敏而容易產生雜訊，而角速度計雖然不容易產生雜訊，但積分所產生的角度卻會隨著時間產生累積誤差，本文使用卡爾曼濾波器融合兩項資訊來即時獲得機器人的姿態傾角，再將姿態傾角分離到各足部方向，隨後再經由逆運動學調整各足部，進而達到恢復姿體平衡的效果。;This thesis applies a fuzzy neural network controller and Kalman filter to control the hexapod for wall following and efficiently adjust the gait to realize stability locomotion. According to the angle position, measured by ultrasonic sensor, between the robot and the wall, the fuzzy neural network controller can control the swing amplitude of the left and right legs of the robot, so that the robot can walk in the complex environment successfully. In addition to walking in an unknown environment, the stability of the hexapod is also a very important theme. The Kalman filter uses an accelerometer and a gyroscope to obtain the real-time robot body attitude, while the tilt angles are separated to the leg directions to change the amplitude by inverse kinematics. Thus, the robot can move forward, and instantly restore horizontal body attitude when walking on oblique terrain. The experimental results show that the method proposed in this research can successfully applied to a real hexapod robot control.
