多機器人協作探索未知區域是多機器人系統重要應用。然而，通訊頻寬資源消耗量、資料安全性、資料一致性、決策共識和機器人異質性等問題限制了多機器人協作探索的效率和完整度。為了解決這些問題，本論文提出了一種基於區塊鏈的多機器人協作地圖探索方法，並以此方法研發了一個多機器人協作探索系統。該系統利用區塊鏈技術實現了機器人間的通訊和資料共享，同時透過身分認證機制建立了機器人間的信任橋樑，並提供了一個公平的分散式決策平台。我們使用一個拓樸地圖作為機器人間的共享地圖，以減少彼此之間大量互動造成通訊頻寬的消耗。在拓樸地圖的基礎上，我們提出了一種輕量化的任務分配演算法Tiny-MinPos，該演算法在先前的研究成果MinPos的架構基礎上進行了擴展，將原本僅適用於佔據網格地圖的限制推廣至拓樸地圖。最後我們在Gazebo模擬器中進行實驗。相較於單機器人系統，我們的系統在區域探索效率方面提升了56%，並在相同時間內的地圖覆蓋率方面平均提升了32%，而Tiny-MinPos相較於Greedy演算法的探索效率提高了15%。實驗結果表明，我們的方法在提高探索效率、增強系統可靠性和實現任務分配公平性方面具有優勢。;Collaborative exploration of unknown areas by multiple robots is a critical application of multi-robot systems. However, issues such as communication bandwidth consumption, data security, data consistency, decision-making consensus, and robot heterogeneity restrict the efficiency and completeness of multi-robot collaborative exploration. To address these issues, this paper proposes a blockchain-based approach for multi-robot collaborative map exploration and develops a multi-robot collaborative exploration system based on this approach. The system utilizes blockchain technology to enable communication and data sharing among robots. Additionally, it establishes a trust bridge among robots through an identity authentication mechanism and provides a fair decentralized decision-making platform. We employ a topological map as a shared map among robots to reduce the consumption of communication bandwidth caused by extensive interactions. Based on the topological map, we propose a lightweight task allocation algorithm called Tiny-MinPos. This algorithm extends the framework of the previous research achievement, MinPos, from its original applicability to occupancy grid maps to include topological maps. Finally, we conducted experiments in the Gazebo simulator. Compared to a single-robot system, our system achieved a 56% improvement in area exploration efficiency and an average of 32% increase in map coverage within the same timeframe. Furthermore, Tiny-MinPos demonstrated a 15% improvement in exploration efficiency compared to the Greedy algorithm. The experimental results highlight the advantages of our approach in enhancing exploration efficiency, improving system reliability, and achieving fairness in task allocation.
