Research On Bandwidth Request And Admission Control For IEEE 802.16 BWA Systems

The IEEE 802.16 BWA (Broadband Wireless Access) system has many advantages, such as easy installation and maintenance, inexpensive operation, connection-oriented service and quality of service guarantee. It is becoming the dominant network access technology in the future.In this thesis, analytic and researching contributions to bandwidth allocation, collision resolution, and bandwidth admission control are shown for the IEEE 802.16 BWA system.The details are as follows.1. According to the IEEE 802.16 standard, substations (SSs) should reserve the required bandwidth before transmitting data packets to the base station (BS). When the SSs share the transmission opportunities to transmit requests, the collision will be inevitable. IEEE 802.16 protocol does not point out the operational mechanism. Most bandwidth allocation algorithms suppose that BS can estimate the number of active SSs,which is not easy in real wireless network. This paper presents a multicast polling mechanism based on the adaptive grouping algorithm. This mechanism brings network traffic analysis and adaptive grouping algorithm in the multicast polling. Firstly, the traffic characteristics such as self-similar, long relative,burst character are analyzed in this mechanism. Secondly, a linear semi-parametric regression model is constituted to evaluate the requesting probability of the non-active SSs.Then the adaptive grouping algorithm based on the virtual Token Ring technology is proposed that operates with the evaluated requesting probability. Simulation results indicate that this polling algorithm can effectively decrease the number of the idle slots and collision slots. 2. A completed contention resolution mechanism includes a pre-contention resource allocation and post-contention collision resolution. The mandatory method of contention resolution that shall be supported by the 802.16 standard is based on a truncated binary exponential backoff algorithm, with the initial backoff window and the maximum backoff window controlled by the BS. With the number of conflicts increasing, the backoff window size is growing rapidly in the algorithm. At the same time, the number of SSs which transmit the bandwidth request relatively keeps the normal level, and there will cause a lot of waste competitive slots. Most algorithm promotion is focus on tuning bandwidth size, for example,MILD algorithm and FCR algorithm. Because of fairness, different types of service flows have an optimal allocation on allowed delay, and retransmitted times of SSs. This article presents a hybrid backoff algorithm which considers both the priorities and fairness. The type of service flow determines the priority and the more times of retransmission hav the higher priority because of fairness. The maximum backoff window and the initial backoff window is controlled based on the fuzzy control method in the next competitive cycle for bandwidth in order to improve high-priority services' quality and ensure fairness between different SS.3. IEEE 802.16 protocol does not define the scheduling methods to guarantee the quality of different services in the BS and SSs.Because that channel quality has a great influence on the modulation and coding scheme in the PHY layer and the transmitting packet size in the MAC layer, a cross-layer scheduling algorithm is proposed in this thesis for the slow time-varying channel in the BWA system. The algorithm adopts the Nakagami channel model. The BS estimates the total allowed bandwidth resources and allocates bandwidth for each active SS according to the signal to noise (SNR) of the channel; SSs again allocate the authored bandwidth according to the QoS of different connections and the adaptive coded modulation and coding methods (AMC) according to SNR.The innovations of this paper include as follows.1. This thesis presents an evaluated algorithm of the bandwidth request of non-active SSs. Firstly, the characteristics such as self-similar, long relative, burst character are analyzed in this mechanism.Secondly, a linear Semi-parametric regression model is constituted to evaluate the requesting probability of non-active SSs based on the evaluated probability and characteristics of service flows.2. This thesis presents a multicast polling mechanism based on the adaptive grouping algorithm.The adaptive grouping algorithm based on the virtual Token Ring technology is proposed that operates with the evaluated requesting probability. Simulation results indicate that this polling algorithm can effectively decrease the number of the idle slots and collision slots.3. This thesis presents a hybrid backoff algorithm which considers about the priorities and fairness.The type of service flow determines the priority and the more times of retransmission are granted the higher priority of an SS because of fairness.In the next competitive cycle, the maximum backoff window and the initial backoff window are controlled based on the fuzzy control method in order to improve high-priority services'quality and ensure fairness between different SS.4.This thesis presents a cross-layer scheduling algorithm for the slow time-varying channel in the BWA system. The algorithm adopts the Nakagami channel model. The BS estimates the total allowed bandwidth resources and allocates bandwidth for each active SS according to the channel signal to noise ratio (SNR); SSs again allocate the authored bandwidth according to the QoS of different connections and the adaptive coded modulation and coding methods (AMC) according to SNR.