This deliverable describes the development of robotic perception algorithms for object pose estimation in HumanTech. Object Pose estimation based on camera images as input is a key-task for localizing objects for robotic grasping. We first provide an overview of the object pose estimation problem and the overall context of the task in the HumanTech project with respect to construction material robotic grasping. Subsequently, we describe the selected object pose estimation framework for the task, the state-of-the-art algorithm ZebraPose developed at DFKI. Finally, we describe the object pose estimation task on the HumanTech object of interest category, construction bricks. We detail the approach for generating and training our machine learning models exclusively on synthetic data and conclude with an evaluation of the brick grasping pose accuracy and the next steps for integration of the method on the human-tech robotic platform for real-time functionality.