Research On Spectrum Resource Occupation Model In Wireless Communication Systems

Nowadays, the use of spectrum resources is far beyond expectations when we consider about wireless communication systems. This phenomenon makes it a rather difficult problem to continue expanding the potential of spectrum. Therefore, how we can make use of wireless spectrum resources efficiently and rationally has aroused world attention.In this thesis, the research is conducted firstly by analyzing the real-time occupation of wireless spectrum resources. On the basis of air interface measurements and spectrum sensing, we could capture primary user(PU) activity from various kinds of wireless communication systems, which plays a combing effect in modeling the usage of spectrum.Real-world spectrum turns to different states when occupation level changes, and a continuous period of time always results in a continuous spectrum states transition. Generally speaking, for the wireless communication systems, it is supposed that all sojourn time of spectrum states approximately follow an exponential distribution, whose stochastic characteristic shows a Markov chain property. Thus, we introduce the modified Markov version, called Hidden Markov Model(HMM), to model the occupation states of wireless spectrum.The typical cellular mobile communication system, GSM, is selected at first as the target wireless network. By measuring the power spectrum density(PSD) of GSM downlink in 4 consecutive weekdays, we seize the opportunity to extract channel state pattern and traffic characters from GSM air interface. The adopting of Gaussian observable Baum-Welch(BW) algorithm provides us a method to analyze the measured power sequences in a HMM-based way, through which we could estimate the channel state parameter set, as well as PU activity of GSM in continuous-time.Then, the popular Wireless Local Area Networks(WLAN) will be studied as another wireless case. The research is started via analyzing the physical layer structure of 802.11 b, including the composition of wireless packet, performance of DBPSK system and the relationship between wireless traffic and occupation of spectrum. Based on these theories, we verify that the real-time measurement data from air interface and the 802.11 b protocol match well under the experimental environment. Demodulation of the measured data packets helps us acquire the packet service parameters in a statistical way. At the same time, according to Poisson distribution of packet departure time, we demonstrate the exponential property of spectrum state sojourn time. As a consequence, we get traffic parameters corresponding to the real packet departure rate by using the HMM estimation method.We verity the high degree of coherency between HMM and wireless spectrum occupation model. On the one hand, HMM helps estimating channel parameter set in GSM downlink, and extracting PU activity. On the other hand, by means of packets demodulation and power adjustment, we calculate the traffic character parameters in both statistical and HMM way successfully in consideration of 802.11 b protocol.