I. Introduction
GAIT biometrics have advantages of recognizing targets at a distance under changing conditions. They tolerates low resolution sensory data and allows for non-cooperative subjects, Conventional video sensor based gait recognition involves large data throughput and high computational cost; infrared video systems are expensive. Advances in sensors and wireless transceivers enable low cost, low data throughput solutions. However, in order to use distributed sensors for gait biometrics, the following challenges need to be addressed:
the increased intra-subject variability of gait attributes in the high dimensional sensory signal space;
the development of stable, repeatable biometric gait feature representations; and
the development of effective feature modeling/testing and data fusion algorithms.