Past Event: CSEM Student Forum

Efficient Approaches for Robust Visibility Assurance

Lukas Taus, CSEM Student

Abstract

Robust visibility assurance is critical in applications such as search-and-rescue, defense, and communications. By addressing multi-coverage objectives where the environment is observed by at least $k$ sensors, we ensure coverage resilience, even in the event of sensor failure. Leveraging greedy algorithms and deep learning techniques, we approximate optimal solutions efficiently, with insights into how data geometry impacts performance. Considering moving agents instead of static sensors introduces a sparse sensing path planning problem, where minimizing sensor use is vital, particularly for scenarios like drones with limited battery capacity. To enhance algorithm efficiency, we propose reconstructing uncertainty-aware paths using neural network predictions, enabling scalable and robust real-time solutions.

Biography

Lukas started the CSEM program in Fall 2021. His current research interests include machine learning, computer vision, and optimal control.

BS, Mathematics, Graz University of Technology
MS, Mathematics, Graz University of Technology