An SQU research team is developing a next-generation robotic system that combines artificial intelligence with autonomous mobility to inspect buildings and navigate across floors without relying on smart infrastructure, IoT connectivity, or human intervention. This work is aimed at improving safety, reliability, and efficiency in building inspection processes, especially in legacy environments.
Funded by the Deanship of Research and College of Engineering, this interdisciplinary project is a collaboration between the Department of Electrical and Computer Engineering and the Department of Civil Engineering at the College. It is led by Dr. Gulam Dastagir Khan as the Principal Investigator, with Dr. Muhammed Bilal Waris as Co-Principal Investigator, and Dr. Taha bin Mubarak Al-Saadi as Co-Investigator.
In the first phase, the team successfully designed and tested a fully autonomous system that enables a quadruped robot to operate standard elevators and navigate multi-floor buildings. Using only onboard components, including a YOLOv11-based vision system, AprilTag-based spatial alignment, and a 4-degree-of-freedom robotic arm, the robot autonomously detects, approaches, and interacts with elevator panels. Real-world experiments in a four-story, non-instrumented academic building demonstrated sub-centimeter accuracy and consistent floor-to-floor navigation.
The second phase of the project aims to extend the robot’s capabilities to perform building inspections. The robot will autonomously assess structural and safety elements such as door frames, fire extinguishers, alarms, sprinkler heads, wall-mounted fixtures, and electrical panels. It will also detect issues like plumbing leaks, insulation faults, and cracks in beams, walls, and ceilings. Integration with thermal and specialised sensors will enhance the range and accuracy of inspection data.
To support intelligent planning and data interpretation, the system will incorporate Building Information Modeling (BIM). This will allow the robot to generate inspection checklists, plan optimal navigation routes, and contextualise sensor outputs with respect to architectural and MEP (Mechanical, Electrical, and Plumbing) layouts, thereby reducing false positives and improving inspection precision.
This research supports the national objectives of the Oman Vision 2040, which calls for smart, resilient urban development through advanced technologies. By reducing human error, improving inspection coverage, and enabling safer operation in hazardous or inaccessible spaces, this project contributes to the future of autonomous infrastructure monitoring in Oman and beyond.