Neuromorphic Vision Chips: intelligent sensors for industrial applications

In the past years researchers at the California Institute of Technology (and in few other research institutions worldwide) have been concentrating their efforts on the design and implementation of analog VLSI neuromorphic systems. Today this emerging technology seems to be mature enough for use in industrial applications. In this paper we give an overview of some of the neuromorphic vision circuits that have been developed at the California Institute of Technology. To point out the advantages and disadvantages of using these types of circuits in industrial applications (with emphasis on the automotive market), we describe a system implemented on a single chip, that detects the direction of heading in the case of translatory ego-motion in a stationary environment. This device contains a variety of analog circuits typically used in neuromorphic systems. We show simulation results applied to vehicle navigation and test data obtained using artificial stimuli in a laboratory environment.