Alexandros Kouris
Imperial College London, University of Patras, Samsung (United Kingdom), Samsung (United States)
Papers
9
Total Citations
204
H-Index
6
About
Alexandros Kouris is a versatile researcher whose work spans autonomous systems, human-robot interaction, and efficient deep learning deployment. His early contributions focused on making robots safer and smarter: his 2016 and 2018 studies on contact distinction in human-robot collaboration — accumulating over 60 citations combined — introduced novel frequency-domain and admittance control methods to differentiate accidental collisions from intentional contact, significantly advancing operator safety. His most-cited work, "Learning to Fly by MySelf" (2018, 92 citations), demonstrated a self-supervised convolutional neural network enabling autonomous UAV navigation without manual labeling, a landmark contribution to drone autonomy. Kouris has also made meaningful strides in efficient AI deployment on resource-constrained hardware, exploring FPGA-based approximate LSTM implementations and on-board DNN optimization for intelligent autonomous systems. More recently, his research on multi-exit semantic segmentation networks addresses the challenge of real-time, adaptable vision on embedded devices — work with growing relevance for mobile robotics and augmented reality. Across these diverse threads, Kouris consistently bridges theoretical innovation with practical deployment, making his research particularly valuable for engineers and scientists working at the intersection of robotics, computer vision, and embedded AI.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3Multi-Exit Semantic Segmentation Networks24 citations · 2022
- 4Contact distinction in human-robot cooperation with admittance control19 citations · 2016
- 5Control of a line following robot based on FSM estimation10 citations · 2018
- 6Approximate FPGA-Based LSTMs Under Computation Time Constraints6 citations · 2018
- 7
- 8Adaptable mobile vision systems through multi-exit neural networks3 citations · 2022
- 9Multi-Exit Semantic Segmentation Networks2 citations · 2021