Artificial Intelligence Robotics and Control Systems Group
Mohamed bin Zayed University of Artificial Intelligence
Current Courses
A PhD-level robotics course on the theory, implementation, and research frontiers of autonomous robot planning and control. The course connects classical motion planning and control with modern AI methods, preparing students to design, analyze, benchmark, and responsibly deploy autonomous robotic systems.
Topics include configuration-space and state-space representations, collision checking, graph search, sampling-based planning, kinodynamic planning, trajectory optimization, model predictive control, safety-critical control, planning under uncertainty, task and motion planning, multi-robot planning, and reproducible robotics experiments. The AI-focused modules cover learned heuristics, imitation learning, offline reinforcement learning, diffusion policies, LLM/VLM-assisted task planning, and vision-language-action policies for robot autonomy.
The course is assessed through a semester-long PhD research project with four checkpoints: proposal, method/prototype progress, experimental evaluation, and final report/demo/oral defense. There are no separate assignments, midterm, or final exam; the emphasis is on original research questions, rigorous evaluation, safety, ethics, and reproducibility.
Course proposal PDF Prescribed texts include LaValle's Planning Algorithms, Principles of Robot Motion, Lynch and Park's Modern Robotics, and Siciliano et al.'s Robotics.
This course introduces the theoretical foundations and practical algorithms for motion planning in robotics. Topics include configuration spaces, graph search methods, sampling-based planning, planning under uncertainty, and symbolic/task planning. Students learn how to design planners with correctness and optimality guarantees and implement them in modern robotics frameworks.
Past Courses
2025
This course explores how robot hands manipulate, sense, and learn from in-hand interactions via model-based control, tactile perception, and data-driven learning. Students program goal-directed, highly dynamic dexterous skills by integrating motion/force control, reinforcement learning, and sim-to-real deployment.
Students implement and test model-based and learning-based in-hand manipulation on a physics simulator and a dexterous hand platform using tactile and visual feedback. Activities include building goal-conditioned controllers and RL policies, sim-to-real transfer, benchmarking on manipulation tasks, and analyzing performance, safety, and robustness.
Student Projects (Selected)
Resources
Software & Tools
- ROS 2 setup guide Course LMS
- Motion capture quickstart Course LMS
- Jetson flashing and profiling Course LMS
- Safety quiz for arena access Course LMS
Templates & Handouts
- Project template Course LMS
- Late policy Course LMS
- Citation and academic integrity guide Course LMS
Course Policies (Common)
- Collaboration: Discuss ideas; submit your own work and credit collaborators.
- Late work: 48-hour window with penalty unless pre-approved (see course page).
- Academic integrity: University rules apply. AI tools allowed only when explicitly permitted; always disclose usage.
- Accessibility: Contact the instructor for accommodations.
- Safety: Complete the Group safety briefing before using hardware.