Juraj Gazda is Vice-Rector for Innovation and Technology Transfer at the Technical University of Košice, where he leads activities in research development, innovation ecosystems, and collaboration with industry. His research focuses on intelligent systems, autonomous mobility, and distributed AI, with an emphasis on real-time decision-making, edge computing, and resilient infrastructures.
He has developed strong international collaborations, including ongoing cooperation with the University of California, Irvine, and research stays at TUHH Hamburg and Ramon Llull University in Barcelona. In addition to his academic work, he has extensive experience collaborating with leading industry partners such as Ericsson and Nokia, contributing to research and development in advanced communication networks and data-driven systems. He is actively engaged in applied research and technology transfer, bridging academia and industry to address challenges at the intersection of artificial intelligence, communication systems, and autonomous technologies.
Presentation Title: Autonomous Driving in Critical Moments: Robust Perception, Multimodal Fusion, and Real-Time Intelligence
Presentation Abstract: What makes autonomous driving truly challenging is not routine operation, but the rare and critical situations that emerge unexpectedly and require immediate, reliable decisions. Addressing these scenarios requires moving beyond average-case performance toward systems that remain robust under uncertainty and adverse conditions. This talk presents approaches for detecting and interpreting such critical events through semantically enriched perception, combining complementary sensing modalities such as RGB and LiDAR to improve both accuracy and efficiency, particularly in low-visibility environments, while also exploring how advanced representations of 3D environments can accelerate visual processing by enabling systems to anticipate observations and operate within real-time constraints. Together, these directions point toward a broader vision of autonomous mobility: systems that are not only accurate, but also efficient, responsive, and dependable when it matters most, achieved through tight integration of perception, communication, and decision-making across distributed and resource-constrained environments.