Particle filtering enhanced human tracking on context-aware robotic system

Abstract

This paper presents the mechanism of visual tracking implemented on a context-aware robotic platform with layered and centralized system architecture. The visual tracking mechanism is developed using Haar-like feature detection algorithm, enhanced by modified Particle Filtering (PF) method, to realize human face tracking and following on a mobile robot platform. Experimental results demonstrate the feasibility and effectiveness of the proposed implementation of visual contexts on the context-aware robotic system.