An approach for movement prediction in Radio over Fiber indoor network at 60 GHz

With its vast unlicensed spectrum of 5 GHz and data speed of up to 2.5 GHz, 60 GHz is envisaged for short range communication in indoor environments. It is driven by the demand for broadband wireless applications such as IPTV, high definition television (HDTV), even uncompressed video. However, in such networks, handoffs are performed frequently due to the small cell (due to smaller range) and the time available for completing a handoff process is short. In order to make a successful handoff, predicting the next location of the mobile user is an important step. It can also enable the system to adapt resources and improve the Quality of Service. In indoor environment, people tend to repeat their movements and also have their selected places, such as offices, libraries, etc., and thus have daily movement patterns. Thus, the system can find their habits from the past. To exploit such patterns, this paper proposes a method using Hidden Markov Model as a learning technique to predict next location of the user. For particular data sets, our experimental results show that the prediction accuracy is up to 81.4% for regular employees and 54.6% for a guest.