Bence Cserna

University of New Hampshire at Manchester

Papers

Total Citations

H-Index

About

Bence Cserna is a researcher whose work lies at the intersection of real-time planning, heuristic search, and autonomous robotics. His primary contributions focus on making on-line planning safer and more efficient for situated systems like mobile robots. In his most-cited paper, "Avoiding Dead Ends in Real-Time Heuristic Search" (2018, 21 citations), Cserna identified a critical flaw in existing real-time search paradigms: they lack a notion of safety, often leading agents into dead ends. He proposed methods to avoid such failures, improving the reliability of concurrent planning and execution. His follow-up work, "Replanning for Situated Robots" (2019, 13 citations), addresses the practical challenge of using low-fidelity world models, which frequently necessitate replanning—a common yet underexplored problem in robotics. More recently, in "Anytime versus Real-Time Heuristic Search for On-Line Planning" (2021, 4 citations), he systematically compared anytime and real-time search approaches, offering insights into their trade-offs for time-constrained AI systems. Cserna’s research is notable for bridging theoretical search algorithms with real-world robotic constraints, making his work highly relevant for students and practitioners in AI planning and robotics.