Date: May 30, 2015 - Half-day Morning
This workshop covers innovative sensing devices, with particular focus on neuromorphic vision sensors, as well as the inference/learning/control methods needed to take advantage of that hardware.
Perception is still one of the main issues in robotics. In the past, leaps in the functionalities of robots were jumpstarted by the availability of new sensor types (the laser range-finder in early 2000s, the 3D cameras in the 2010s). So what will be next enabling technology in sensing?
Topics of interest include the emerging field of neuromorphic vision sensors (DVS, ATIS, etc.), which might be the technology enabling high performance yet low power small UAVs, as well as other unconventional vision systems (4D light field/polydioptrical cameras, template-based optical computing) and other currently under-utilized sensor modalities.
The program a as a whole will provide a coherent tutorial on:
- neuromorphic vision sensors
- spike-based and AER-based algorithms
- neuromorphic computation substrates (including memristors)
- neurobiology of the insects' neural circuits dedicated to visual processing
|8:45‑9:15||Christian Brandli (Insightness)||"Event-based Vision Sensors: Principles and Potential of a Neuromorphic Technology"|
|9:15‑9:45||Jörg Conradt (Technische Universität München)||"Application of neuromorphic sensors to the robotics scenario"|
|9:45‑10:00||Michael Milford (QUT)||"Place Recognition with Event-based Cameras and a Neural Implementation of SeqSLAM"|
|10:30‑11:00||Andrew Straw (Institute of Molecular Pathology, Vienna)||"Neural circuits for visual processing in the fruit fly"|
|11:00‑11:25||Mirko Prezioso (UCSB)||"Operation of a memristor-based single layer neural network"|
|11:25-11:50||Chiara Bartolozzi (Italian Institute of Technology)|