目前機器人的發展,較多都是輪型機構為主。在高低差不大的地面上,此類的機器人不論在轉彎或是直行,都較非輪型之機器人靈活且快速。但如何使機器人利用高行動力執行多項任務,為本論文研究的課題。 本論文之輪型機器人為附載能力較強的四輪架構。前輪以高扭力直流馬達驅動,而為了減少迴旋半徑,以全向金屬球輪作為後輪使用。利用FPGA晶片作為處理核心控制周邊硬體裝置,賦予機器人行動能力以及感官。 避障系統使機器人在未知的環境下,可以安全地移動。因此,本論文採用了模糊避障系統。有鑑於一般的模糊避障系統,並不會紀錄過去的反應狀態,導致避障功能之效率較低。因此,本論文將上一次的反應狀態作為模糊控制系統之輸入,提高避障效率。人物跟隨系統利用類哈爾特徵(Haar-like)進行特徵擷取以及匹配,追蹤已知的標誌。藉以達到即時辨識的能力。透過位置以及大小,估算使用者所在方向以及距離,作出相對應的動作。平滑路徑規劃系統,利用了跳點搜尋()以及A*演算法。再以貝茲曲線模型,重建出平滑路徑。透過遠端遙控系統,使用者可以透過手機APP對機器人下達指令以及遙控。 實驗結果顯示,模糊避障系統、人物跟隨系統、路徑規劃系統以及遠端遙控系統,成功應用在Altera公司的FPGA開發板上。此外,在模糊避障系統中,比一般模糊避障系統有效率的避開障礙物。
In these days, the wheeled robots are popular in the robotic search area. The wheeled robots can move more agilely in smooth ground. With the great mobility, the wheeled robots could complete more difficult missions. In the thesis, the structure of the wheeled robots is the acrylic plane with four wheels which can improve the carrying capacity. The high-torque motors force the front wheels. To reduce the radius of gyration, the ball casters with omni-direction metal are utilized as the rear wheels. The FPGA is employed as the core of the robot and it control all of the peripheral devices in the robot. In addition, the multi-function circuits also are built up in the FPGA. In fuzzy obstacle avoidance, it takes the last state as additional input which could make the robot efficient in obstacle avoidance. People following system is adapted with Haar-like feature which can execute feature extraction and feature matching in real-time. The distance and the direction of the people could be estimated by the area of the mark detected. In Smooth Path Planning system, Jump point search and A* algorithm is exploited. After the nodes of the path is found out, they are connected by the rational bézier curves model in path reconstruction. Remote controlling system make user can control the robot by mobile phone. The experiment shows the fuzzy obstacle avoidance system, people following system, smooth path planning system and remote controlling system are implanted in Altera FPGA demonstration board. Additionally, the proposed fuzzy obstacles avoidance system can escape from obstacles more quickly than the general design.