本研究將針對機器人同時定位、建圖、與移動物體追蹤(SLAMMOT)問題及其 在室內動態環境中之應用,進行為期兩年的研究。研究的議題包括移動物體偵測、影 像特徵偵測與描述、以及在未知的動態環境中巡航等。目標是在兩年內完成即時視覺 式SLAMMOT 演算法,可以應用在智慧型行動機器人系統上,使機器人在未知的動 態環境中具備同時定位與建圖的能力,以順利執行巡航任務。 本計畫分兩年執行:在第一年,將提出動態環境中移動物體偵測的演算法, 也將改良影像特徵偵測與描述方法,以及整合應用於手持攝影機或行動機器人,以便 在未知的動態環境中執行同時定位與建圖的任務。本年度將使用擴張型卡爾曼過濾器 (EKF)進行系統狀態的估測,因此,也將探討系統具備未知輸入項的問題與EKF 效能 改善等議題。第二年將利用稀疏擴張型訊息過濾器(SEIF)的概念來設計機器人的 SLAMMOT 演算法, 搭配區域地圖的管理, 規劃在室內長距離的範圍進行 SLAMMOT。最後,將實現EKF 與SEIF 兩種型式的即時視覺式SLAMMOT 演算法, 並應用在機器人系統於未知的動態環境中同時定位與建圖的任務。本計畫所發展的機 器人同時定位、建圖、與移動物體追蹤,將以實地測試方式評估所提相關演算法的實 用性。 A two-year investigation plan is proposed in this research to focus on robot simultaneous localization, mapping and moving object tracking (SLAMMOT) and its application in indoor dynamic environments. The research topics include moving object detection, image feature detection and description, and robot navigation in unknown and dynamic environments. The aim of this project is to develop real-time vision-based SLAMMOT algorithms for intelligent mobile robots, providing the robots with suitable localization and mapping competence in unknown and dynamic environments in order to perform the navigation tasks. This project will be performed in two years. In the first year, a novel moving object detection algorithm as well as a modified image feature detection and description procedure are proposed for a hand-held camera or a mobile robot in order to perform localization and mapping in unknown and dynamic environments. The SLAMMOT is implemented by utilizing an extended Kalman filter (EKF) for estimation of system states. Therefore, the system with unknown inputs will be discussed and the performance of EKF will be analyzed. In the second year, the concept of sparse extended information filter (SEIF) is applied to design the SLAMMOT algorithm for mobile robots. Integrated with a local map management strategy, the SLAMMOT algorithm will be able to perform the tasks of localization and mapping in long-distance and large-area environments. Two real-time vision-based robot SLAMMOT algorithms will be implemented, one based on EKF and the other on SEIF, and be further applied on robot localization and mapping in unknown and dynamic environments. The developed algorithms will be integrated with real robot systems, and furthermore the integrated system will be empirically validated based on experiments to evaluate the practical usage of the developed algorithms.