Performance estimation of oversampled low bit depth, bio-inspired motion detection system

Abstract

The Reichardt Correlator is a bio-inspired motion detector with its low computation complexity and is commonly used for velocity estimation of objects. Traditionally, the input signals to the Reichardt Correlator are represented at high bit depths and sampled at Nyquist rate. With advances in digital microelectronics and mixed signal techniques, this paper presents a new approach for using oversampled, low bit depth representation for velocity estimation using Reichardt Correlator. This is achieved by considering the trade-off between oversampling ratio and bit depth, commonly found in oversampled data converters. The presented results show that the proposed approach offers good performance, with a potential for reduction in hardware complexity.