Abstract As robots become increasingly integrated into human environments, the ability to accurately perceive and interact with three-dimensional objects is crucial for their effective operation. This thesis explores the fundamental challenge of 6D object pose estimation, a critical component in enabling robots to understand and manipulate the diverse objects in their surroundings. The concept of 6D pose, encompassing both 3D position and 3D orientation, provides a complete description of an object’s spatial configuration. Accurate 6D pose estimation is essential for a wide range of robotic applications, including grasping and manipulation, augmented reality, autonomous navigation, and human-robot collaboration. While humans effortlessly perceive and interact with objects in 3D space, replicating this capability in robotic systems remains a significant challenge due to factors such as variable lighting conditions, occlusions, and the diversity of object shapes and textures. In this thesis, we address the limitations of current 6D pose estimation methods through four main contributions: First, we propose novel approaches that leverage category-level knowledge to estimate the poses of previously unseen objects. Second, we introduce UDA-COPE (Unsupervised Domain Adaptation for Categorylevel Object Pose Estimation) to enhance accuracy and robustness in real-world scenarios. Third, we present TTA-COPE (Test Time Adaptation for Category-level Object Pose Estimation) for more practical, adaptive pose estimation. Fourth, we explore methods that go beyond category-level estimation to handle any object without prior category knowledge. Our research aims to bridge the gap between human and machine perception in the 3D world by advancing these areas while extending robotic capabilities beyond human limitations. Our work focuses on developing more robust and generalized approaches to object pose estimation, enhancing the ability of robots to perceive and interact with their environment in increasingly complex and dynamic scenarios. The proposed research has significant implications for enhancing machine-based 3D spatial understanding, enabling automated systems to assist in daily tasks, provide care, and collaborate effectively in manufacturing and home environments.     Keywords 6D Object Pose, Novel Object Pose, Robot Manipulation, Augmented Reality