Mohamad Jaber

Adaptive execution of robot skills using computer vision – Built a custom 2D object detector using TensorFlow Objection Detection API. – Exported RGB and mask images along with their corresponding bounding cuboids as labels using NDDS. – Trained the singleshotpose model by Microsoft using the annotated data to predict the 6D object poses. – Developed a virtual camera using C# in RobotStudio. – Validated the trained model in RobotStudio by achieving an End-to-End process for a sorting application.

– Conducted a case study technical report on a data science topic. – Recognized the importance of vision in robots as an industrial necessity and formulated a topic based on it.

– Extracted geometrical and topological features from CAD files using Open CASCADE. – Calculated and configured the forward kinematics of a Reis robot in MATLAB. – Programmed the inverse kinematics of a Reis robot in C# and MATLAB and assured consistency between joint and task space. – Supported the team in many experiments, research studies and presentations in the fields of robotics and computer vision.

– Modeled and verified a state-space representation based on a heat dynamics system of a building which was implemented in Python using control theory knowledge. – Integrated the developed model with an extended MILP optimization problem using Python package (Pyomo). – Documented the developed model into a research paper using LaTeX. – Analyzed and visualized energy generation/consumption data using Python packages (NumPy and Pandas)

Machine learning, Control theory, Robotics, Data Science

Control theory, Statistics, Mechanics, Electronics