Robotics + GPU
Action, language, and acceleration
PhD work connecting iCub humanoid learning, action structure, neural networks, CUDA, Aquila, and ESA rover-control research.
Robotics workPersonal legacy archive
Martin Peniak
Robots, cameras, coordinates, and place.
A life spent making ideas answer to the physical world.
This is a personal archive of work that stayed close to the physical world: Plymouth, iCub humanoid robotics, Aquila, NVIDIA CUDA, ESA rover simulation, TEDx Bratislava, applied AI cameras, spatial intelligence, Project Tao, and the Legacy Writing Archive.
TEDx Bratislava
My dream about AI.
A public telling of the early story: childhood fascination, restart, study, and the long route into robotics.
Watch TEDxPublic records
Records that anchor the story.
These links stay close to hand so the public record is easy to follow, without turning the archive into a wall of credentials.
Work
A map of the work.
ESA rover simulation, iCub learning, CUDA experiments, edge cameras, spatial calibration, and Tao: each chapter tests perception against something physical.
Applied vision
AI cameras and edge vision
Camera-as-computer prototypes, edge inference, synthetic worlds, and the pressure of making ideas work outside the lab.
Camera work
Spatial intelligence
From pixels to place
Multi-camera calibration, floorplane reasoning, uncertainty, topology, and the work of making observations belong to the same world.
Spatial workSources
Primary sources, plainly linked.
Thesis, papers, code, applied-vision articles, spatial artifacts, talks, profiles, and older links live together here.
Tao
A place built slowly.
Tao is where the record becomes land, water, paths, structures, gardens, seasons, memory, and care.
Over years, rough land became water, paths, planting, structures, and care.
Lineage
Robots, cameras, coordinates, place.
The tools changed from rover simulators and humanoid robots to camera systems, mapped spaces, and land. The recurring question stayed concrete: what is here, where is it, and what still works after contact with the world?
ESA rover
Autonomy and sensing against terrain, uncertainty, and planetary-robotics constraints.
iCub and Aquila
Action, language, neural dynamics, body constraints, and GPU-accelerated experiments.
NVIDIA and CUDA
Early large-scale GPU robotics work made visible through CUDA records, GTC, and public talks.
Applied cameras
Perception systems moved outside the lab into edge devices, workflows, and real-world pressure.
Spatial intelligence
Cameras become more useful when observations share coordinates, topology, and uncertainty.
Tao
A place chapter: land, water, paths, structures, seasons, repair, and memory.
Videos
Talks, demos, and build films.
Talks, demos, and films give the record movement: research, applied vision, synthetic worlds, and Tao becoming real over time.
