A clustering-based context-aware mechanism for IEEE 802.21 Media Independent Handover

The IEEE Media Independent Handover (MIH) services are not aware of different user contexts, and thus cannot provide context-aware services to improve user experience. To address this problem, we extend MIH framework with a context-awareness module. More specifically, we present a clustering based context-aware mechanism for predicting stability of wireless links in our framework. Unlike existing works, which require prior knowledge on network environment and user mobility pattern, our proposed mechanism can automatically be aware of environmental interference, user location, user handover policy, and user mobility. A K-mean clustering method is adopted to process the classification of user context, and then the link stability is predicted with conditional probability according to different classifications. Experimental results show that our proposed context-aware prediction can well guide the mobile user to obtain a stable wireless link.