In personal area network (PAN), high transmission rate and low design complexity in medium access control (MAC) protocol are the benefits that wireless local area networks (WLAN) possess. Hot spot cells and indoor environments widely apply it for diverse applications. Due to channel sharing, the MAC protocol is the key role to determine the efficiency and performance of the WLAN. The Abramson proposed an elegant MAC protocol, called ALOHA [1]. In ALOHA, high collision probability result form stations allowed transmitting immediately upon receiving data from upper layers.
To decrease the collision probability, carrier sense multiple access (CSMA) scheme [2]
requires stations to transmit until the medium becomes idle. When a station detects the channel is idle, it can transmit with a probability of 1 or p (0 < p < 1). The former is called 1-persistent CSMA and the latter is p-persistent CSMA. In the IEEE 802.11 [3], it adapt the CA scheme of DCF further reduces frame collision probability by requiring each backlogged station to perform binary exponential backoff (BEB) after the medium becomes idle. In BEB, if a station successfully transmits a frame, its contention window will be reset to an initial value. However, if the transmission fails, the window size is doubled. Nevertheless, owing to the situation that collided stations would have smaller probability to access the medium than new stations, it generally entails larger delay variance on stations and results in unfairness.
The important and challenging issue is to divide the channel fairly among stations.
Therefore, the design of efficient MAC protocols with high-throughput performance, as well as a high degree of fairness performance, has been a major focus in WLAN
research areas [4], [5], [49]-[51]. There have been many studies of backoff algorithms [6]-[13], [16], [45], but all of them did not address the problem of larger delay variance.
For real-time applications, higher delay variance leads to larger amount of dropped packets due to excess delay. For non-real-time data applications, on the other hand, higher delay variance usually causes larger requirement of storage buffer or more probability of buffer overflow. Hence, in order to reduce the delay variance, the radio resource over WLAN interface is necessarily to be fairly shared by an effective MAC protocol. The fairness problem of MAC in WLAN that some algorithms solve was proposed [14], [15]. However, channel throughput is decreased because of the high collision probabilities.
Dynamic contention window (CW) schemes [16]-[18], different maximum packet length scheme [18], and various interframe space (IFS) schemes [18]-[20] are usually adopted to design the priority differentiation in order to support multimedia services for the IEEE 802.11e [52] WLAN. However, owing to the backoff scheme, large delay variance in the same access category (AC) would still be arisen by these solutions.
Obviously, larger probability of quality-of-service (QoS) violation of multimedia traffic is brought about by higher delay variance because of excess delay.
In WMANs, the problem of future wireless communication is resolved by multiple-input multiple-output (MIMO) based orthogonal frequency division multiplexing (OFDM) because it helps achieve high system capacity and provide transmit/receiver diversity for reliable communication link. Downlink resource management for multiuser OFDM (MU-OFDM) systems has recently been investigated [21]-[25], in which topics were emphasized on transmission power allocation, subcarrier allocation, bit allocation, or adaptive modulation and coding (AMC). The goal of the
fairness, or guarantee QoS requirements.
Many papers investigated the downlink resource allocation [21]-[32] but few papers probed into the uplink resource allocation. Both downlink and uplink perform the resource allocation primarily through the base station (BS). The power distribution over the selected set of subcarriers for every user is included in the algorithm in [35] so that it minimizes the total power being used. In [36], a greedy subcarrier allocation algorithm, based on a marginal rate function, and an iterative water-filling power allocation algorithm were proposed. A practical algorithm and the optimization problem were presented in [37]. All of them can nearly reach an optimal solution but they did not focus more on the QoS requirement. The power saving in IEEE 802.16 OFDMA systems via an efficient uplink resource allocation was shown in [38]. While guaranteeing BER, it minimizes the required transmissions power through adaptively adjusting the modulation and coding scheme. However, they don’t attend to their differentiated QoS requirements and multiple services. [39] exhibited an efficient and fair scheduling (EFS) algorithm for each time slot in IEEE 802.16 OFDMA/TDD system. A fixed priority scheme which gives priorities to service traffic according to their QoS requirements is applied to design the EFS algorithm.
The bandwidth is allocated according to channel quality and queue state of the traffic for SS with real-time and non-real-time polling services in [40]. From the previously mentioned works, people either omitted or simplified the QoS requirements and fairness issues. The QoS requirement usually refers to a predefined weight which corresponds to the fixed priority scheme or a minimum required transmission data rate.
However, the design of radio resource allocation for practical applications should include the delay bound and the packet dropping rate, regarded as essential QoS requirements to provide multimedia real-time traffic. In addition, buffer conditions of
different traffic types and realistic traffic models should be taken into account.
For MIMO-OFDM systems, the exponential increases in the computational complexity on radio resource scheduling for downlink multiuser is proportional to the number of subcarrier, multiuser, transmitting antenna, and receiving antenna. The multiuser scheduling algorithms for system throughput maximization with reduced complexity in a downlink MIMO/OFDMA system were proposed [26]-[29]. They decoupled the multiuser scheduling problem into frequency and spatial domains.
Multiple parallel independent single-user MIMO channels are decomposed from the multiuser downlink MIMO channel by the preprocessing scheme. However, the number of transmitting antennas restrains the number of simultaneously transmitted users.
Computational complexity of the scheduling algorithm is still too high and the QoS requirements and user demand were not considered in the scheduling algorithm.
A fixed priority algorithm was proposed [30] in relation to the QoS requirement in MU-MIMO-OFDMA system. The non real-time traffic’s rate of transmission is too low to fulfill the requirement rate while the real-time traffic can be provided in time at low traffic intensity. A dynamic priority scheduling scheme was proposed in [32]. With that scheme, not only is high priority given for urgent users but also the priority of users is dynamically adjusted frame by frame. However, the ARRA does not give the clear differentiation of real-time service from non-real time one but depends on the time to expiration while adjusting the priority.
This is also an important issue: the tradeoff between system performance and computational complexity. Optimal solution [41] can be achieved by the greedy algorithm which performs symbol by symbol allocation, but it causes high computational complexity. The symbol-by-symbol allocation algorithm costs high
defined in IEEE 802.16 for downlink and uplink, respectively. Moreover, most resource allocation algorithms are not only designed for downlink but also claimed to be compatible with uplink. Even so, both the uplink frame structure (UL-MAP) and downlink frame structure (DL-MAP) have different definitions in IEEE 802.16 specifications [42]. Therefore, to meet its individual frame structures, a design of an efficient and feasible resource allocation algorithm for either downlink or uplink is particularly needed.