| 摘要: | 近年來，隨著人與機器人之間互動功能日漸頻繁，顯示了寵物機器人逐步進展成為下一代電子玩具，而統計研究指出產品的創新性有助於提高消費者的接受度，未來寵物機器人將具備更多的可變性與趣味性，為了融入人類的生活而不顯得冰冷生硬，寵物機器人需更加擬人化。然而機器人的追蹤系統就顯得非常重要，有不少研究提出使用距離感測器，並以智慧演算法為基礎的控制器，設計出多種控制方法，強化機器人的即時追蹤的速度及精確度。
本論文在設計一適用於寵物機器人的TSK模糊移動控制器，並以基因演算法做為參數最佳化的調變機制。設計過程，首先建立TSK模糊規則庫，再以基因演算法(Genetic Algorithms, GA)的複製，交配與突變運算，驅使誤差適應函數最小化，以找出規則庫的最佳參數，然後代入TSK模糊控制器中，進行線上控制，為加快運算速度，本論文採用實數型的染色體。
為驗證所提方法之有效性，研究過程將此控制器以MATLAB模擬，實驗室的真實寵物機器人平台實驗，從模擬與實驗結果來看，提高了寵物機器人即時追蹤能力。
In recent years, with the interaction between people and robots increasingly frequent, showing the pet robot gradually progress into the next generation of electronic toys, the future of pet robots will have more variability and fun, pet robots need to be more anthropomorphic. However, the robot tracking system is very important, there are many studies proposed the use of distance sensors, the use of intelligent algorithm based on the controller, design a variety of control methods to enhance the robot's instant tracking speed and accuracy.
In this paper, we design a TSK fuzzy mobile controller suitable for pet robots, and the genetic algorithm as a parameter optimization of the modulation mechanism. Design process, the TSK fuzzy rule base is established firstly, and then the genetic algorithm is used to minimize the error adaptation function to find out the best parameters of the rule base, and then substitute TSK fuzzy Controller, the online control, in order to speed up the operation speed, this paper uses real type of chromosome.
In order to verify the effectiveness of the proposed method, the research process of this controller to MATLAB simulation, laboratory real pet robot platform experiment, from the simulation and experimental results, pet robot has a good real-time tracking ability. |
