Rethinking robot navigation with more prediction instead of planning
Aykut I艧leyen defended his PhD thesis at the Department of Mechanical Engineering on November 3rd.
Autonomous mobile robots are becoming increasingly important in everyday life, from moving goods in warehouses to assisting elderly people at home. For these robots to be truly useful and reliable, they must move safely and efficiently, even in cluttered or partially unknown environments. However, most current systems rely on frequent replanning cycles, which are often slow and computationally expensive. In his PhD research, Aykut I艧leyen proposes a new way of thinking to drive robots to deal with these challenges.
Aykut I艧leyen鈥檚 research introduces a paradigm shift: 鈥減lan less, predict more.鈥� His research demonstrates a new method that allows robots to move safely by continuously checking the safety of their predicted motion, without needing to stop and replan. This allows robots to respond more quickly and intelligently to their surroundings, without the need for complex recalculations every time something changes.
Safer and more efficient robots
I艧leyen focuses on the unicycle robot, a common type in robotics. He demonstrates how his research can be used for safe navigation, motion planning, and proactive exploration in unknown spaces for this type of robot. The proposed methods were validated through simulations and real-world experiments and they show promising results for making robots safer, more efficient, and better suited for real-life applications. This has the potential to improve the safety and scalability of future robotic systems across logistics, healthcare, and public spaces.
Title of PhD thesis: . Supervisors: Dr. 脰m眉r Arslan, Prof. Ren茅 van de Molengraft and Prof. Nathan van de Wouw.