Evaluation of an artificial CMOS retina sensor for tracking systems

Smart vision chips will be the basic component Of future intelligent systems. Conventional vision systems, using a CCD camera and a digital processor, have mainly been limited to specific industrial and military applications due to the cost, size and complexity. Smart vision chips, which include both the photosensors and elements of (analog and digital) parallel processing, have been under research for more than a decade and illustrate promising capabilities. The system we use is based on an artificial multi-mode programmable retina sensor. This retina is a CMOS image sensor developed at IEF (Institut d'Electronique Fondamentale, Universite Paris XI, Orsay-France), and includes some silicon analog and digital image processing algorithms. We describe a system, consisting of a Minicar platform and one or two CMOS retina sensors. Specifically, we demonstrated how the hardware capabilities of the retina sensors can be used in enabling the system to reliably and autonomously track a laser point. We present the data issued from the retina sensors and we focus on the analysis of the hardware-software co-designed system.