雲端上的編舞家：以 生成式AI 實現「無人機群飛」的座標自動化與跨域實踐
講者/Speaker： 林木盛, Lin, Mu-Sheng
無人機群飛展演雖具備極高的 STEAM 教育潛力，但在實際教學現場，繁瑣的幾何座標換算與實機操作的碰撞風險，往往消磨了學生的學習動機，也讓教師在有限課時內難以兼顧理論與實務。為突破此教學瓶頸，本研究建構並實施了一套結合「生成式 AI、GeoGebra 動態數學驗算、ASST 圖形化程式模擬」的跨領域教學架構。本方案將 AI 精準定位為「座標轉換協作支架」而非單純的解答機，引導學習者落實「生成—驗證—修正」的閉環實作機制。
課程依循任務梯度，由單機操控循序漸進至方陣、圓形乃至心形等多機編隊設計。同時，我們將 AI 風險控管內化於教學流程中，要求學生對模型產出的數據進行批判性檢核與邏輯除錯。經 180 名修課學生實證，本課程不僅成功產出高重現性的群飛配置，學生在空間建模、計算思維及跨域問題解決等綜合能力上亦有顯著成長。研究結果證實，具備教學意圖地導入生成式 AI，能大幅降低前沿科技的學習壁壘，為技職教育體系提供具實務參考價值的創新教學模式。
While drone swarm performances hold immense potential for cross-disciplinary STEAM education, the tedious nature of geometric coordinate calculations and the collision risks associated with physical flights often frustrate students and overwhelm instructors working within limited class hours. To overcome this pedagogical bottleneck, this study developed an integrated instructional framework combining Generative AI, GeoGebra for dynamic mathematical verification, and the ASST graphical simulation platform. By deliberately positioning AI as a "coordinate translation scaffold" rather than a mere answer generator, the curriculum guides learners through a closed-loop "generate-verify-correct" mechanism.
The task-driven course progresses systematically from basic single-drone operations to complex multi-drone formations, including squares, circles, and heart shapes. Furthermore, AI risk management is intrinsically embedded into the workflow, explicitly requiring students to critically validate and debug AI-generated data. Empirical results from 180 enrolled students indicate that the framework not only yielded highly reproducible swarm configurations but also significantly enhanced students' spatial modeling, computational thinking, and metacognitive problem-solving capabilities. These findings confirm that purposefully integrating generative AI as an instructional scaffold effectively lowers the entry barriers to advanced technologies, offering a robust, innovative paradigm for technical and vocational education systems.
- 論文全文：雲端上的編舞家：以 生成式AI 實現「無人機群飛」的座標自動化與跨域實踐
- 影像： 相關影片