This talk introduces AI-based robotic applications developed for higher education and shows how building these systems can make AI and robotics concepts tangible through problem-based learning (PBL). Building on key outcomes of the European AiRobo project, the talk outlines practice-based learning designs that bridge abstract concepts and working robotic systems, highlighting both educational and socially impactful applications, including assistive technologies for people with special needs.
The presentation is structured around modular robotic case studies that have been successfully developed as course projects, capstone topics, and staff-training exemplars. Each case study is mapped to core theoretical content—AI algorithms, programming paradigms, robotics foundations, and formal verification—so that students can see why the theory matters, how it is implemented, and how practical constraints shape the final design.
  1. Indoor navigation with Pepper and Scout robots. By combining a social robot (Pepper) with a mobile robot (Scout), this case study supports learning activities around localization, path planning, and human–robot interaction. It links foundational methods in graph search, state estimation, sensor fusion, and safe behavior design, while exposing practical constraints such as noisy perception, dynamic obstacles, and real-time decision requirements.
  2. 2D/3D mapping with AgileX Scout Mini and LiDAR. This case study introduces SLAM principles and robotic perception through hands-on map generation and evaluation. Students work with coordinate frames, point-cloud processing, occupancy grids, and basic calibration, and they practice interpreting quantitative metrics (coverage, drift, loop-closure quality) and use these measures to iteratively improve the system design.
  3. An AI-based Tic-Tac-Toe robotic arm. A compact manipulation scenario is used to connect computer vision, motion planning, and game AI. The pipeline integrates camera-based board detection, rule-based or search-based decision logic, and pick-and-place control. The case study also enables discussions on verification-oriented thinking: specifying critical properties (e.g., legal moves, collision avoidance) and validating them through systematic testing or formal specifications.
  4. Building a telepresence robot. This project emphasizes systems integration: networking, streaming, remote control, autonomy assist, and interface design. It supports instruction in software architecture, concurrency, and evaluation with users, highlighting the engineering practices required for robustness and reproducibility beyond prototype-level demonstrations.
  5.  A 3D-printed rehabilitation exoskeleton. This case study shifts the emphasis to responsible robotics, including human-in-the-loop control, safety constraints, ergonomics, and ethical considerations. It illustrates how biomechanical requirements can be translated into design parameters, how prototypes can be validated, and how limitations and risk mitigations can be documented and communicated.
Across these examples, the talk highlights the PBL workflow: starting from a real life problem, identifying the concepts and methods required to address it, implementing and integrating solutions, and evaluating outcomes through measurable criteria and reflective reporting. All implementation code developed within the AiRobo project is freely available and can be reused or adapted for teaching and research.