Subcarrier allocation in conventional and network coded OFDMA systems

In wireless communications systems, the fading effect, one of the basic factors which degrades the performance, is an important problem for reliable and qualified transmission. Thanks to channel coding and diversity techniques which are used to reduce fading effect. Cooperative diversity is one of the diversity techniques and by using this technique, error performance is improved by sacrificing the data rate. It is a solution for the increasing number of users and the increasing demand for high quality wireless services. The cooperation among mobile users of a wireless system has the potential to provide an increased capacity in comparison with the systems without using cooperation. Network coding is a technique developed to avoid the loss of data rate for cooperative systems and there have been many ongoing researches about this topic in recent years. It is a useful technology that can increase the total throughput of wireless networks and allows intermadiate nodes in networks to encode several received packets into a coded packet before forwarding. In random network coding that is a more advanced version of network coding, data packets are encoded by multiplexing with random generated coefficients. Thus, random network coding is more robust against the packet loss and latency. At the same time, transmission rate and bandwidth efficiency of the systems are improved by using this technique. In wireless communications, a problem which is closely related with transmission rate and bandwitdh efficiency is resource allocation problem. Basic principle of resource allocation is to maximize the system capacity. By dynamically allocating resources to users in a system, system capacity maximization can be done. Therefore, it is important that restricted frequency band must be used by users effectively as much as possible. OFDM is a multiplexing technique which subdivides the available bandwidth into multiple orthogonal frequency subbands and has an importance for existing systems and new generation communications systems due to its ability to improve the transmission efficiency and reliability in frequency selective fading channels. OFDM systems are able to maximize spectral efficiency without causing adjacent channel interference. One of the important applications of OFDM for multi-user communications is OFDMA and it is a multiplexing scheme that provides multiplexing operation of user data onto the subcarriers. Especially in the multi-user systems, available sources must be used optimally under bandwidth limitations. The source allocation comes up as a more important problem in recent years due to increasing number of users who use the mobile devices. In OFDMA systems, user data packets are transmitted via subcarriers. Available capacity is used effectively by allocating subcarriers to the users optimally. By this way, outage probability of users can be minimized. In an OFDMA system, if instantaneous channel capacity is lower than a target transmission rate due to frequency selective slow fading, subcarriers may be in outage. Besides, if all subcarriers for a user are in outage, that user will also be in outage. So, allocating non-outage subcarriers to users is an important problem for OFDMA systems. In order to solve this allocation problem, an OFDMA system can be represented by using random bipartite graph model which is a subject of graph theory. Random bipartite graphs can be used to analyse the multi-user subcarrier allocation problem over frequency selective slow fading channels. In random bipartite graphs, subcarriers can be allocated to users by employing maximum matching algorithms. This minimizes the user outage probability. Moreover, each user have the same opportunity in terms of subcarrier allocation. In other words, the fairness between the users is guaranteed. RVRHK algorithm developed based on the Hopcroft-Karp algorithm can be used for subcarrier allocation with maximum matching in random bipartite graphs. According to the results obtained in the literature, diversity gain can be achieved via maximum matching algorithms. RVRHK algorithm takes advantage of the maximum matching technique to resolve the user conflicts in OFDMA systems. In addition, the channel state information needed by this algorithm is only 1 bit/subcarrier/user to resresent the outage scenario. This algorithm allocates only one subcarrier to each suitable user. For more general cases, H-matching approach is utilized. An improved version of RVRHK algorithm is R^2 EHK algorithm where ℋ -matching approach is utilized. In order to guarantee the fairness between users, a randomly rotation operation is performed among users in random bipartite graph for both RVRHK and R^2 EHK algorithms. By using H-matching approach in R^2 EHK algorithm, equal number of multiple subcarriers can be allocated to users in an OFDMA system as long as each user which is not in outage has same number of subcarriers. In this thesis, the performances of RVRHK and R^2 EHK algorithms are investigated for Nakagami-m channels for the first time in literature. The obtained results show that diversity order of the OFDMA systems increases with m parameter for Nakagami-m fading channels. By this way, outage probability is also improved. Besides, Multi-R^2 EHK algorithm is developed in accordance with requirements of users. Indeed, in subcarrier allocation solutions via random bipartite graphs, it is generally assumed that user transmission rates are fixed. In case of users in an OFDMA system utilize different services, they may have different bandwidth requirements. So, the number of subcarriers to be allocated to users and the user transmission rates may be different. Therefore, different numbers of subcarrier allocation problem is examined under two headings: fixed bandwidth efficiency per user and variable bandwidth efficiency per user. For the first topic, a fixed bandwidth efficiency is promised for each user, but they have different numbers of subcarriers. So, the user transmission rates will be different according to number of subcarriers allocated. For the second topic, the bandwidth efficiencies of the users will be variable. In this thesis, a new version of R^2 EHK algorithm, which allows different numbers of subcarrier allocation to users, is developed. The proposed algorithm uses the H-matching approach. By using this new algorithm, different numbers of subcarrier allocation can be done optimally according to the user requirements at the cost of a degradation in the outage performance. The results obtained for fixed bandwidth efficiency by using Multi-R^2 EHK algorithm show that the outage probability of users with more number of subcarriers is lower than the others. In addtion, it is shown that the users with less number of subcarriers are dominant in determining general outage performance of the system compared to the users with equal number of subcarriers. If the bandwidth efficiency is assumed to be fixed, when a user in the system has less number of subcarriers than the others, its user outage performance becomes worse. It is also observed from the results obtained for variable bandwidth efficiency by employing Multi-R^2 EHK algorithm, a users with more number of subcarriers should satisfy a higher transmission rate. In this case, the outage probability of this user will be higher than other users. In addition, if the bandwidth efficiency for each user is changed proportionally with the number of subcarriers, all users will have same outage performance. In the last chapter of this thesis, the analyses given for traditional OFDMA systems are generalized to network coded OFDMA systems. A new bipartite hypergraph model is proposed to represent the connections between users and user/relay subcarriers in the network coded OFDMA systems. As in aforementioned techniques, subcarrier allocation can be done via bipartite hypergraphs by minimizing the user outage probabilities. The proposed Multi-R^2 EHK algorithm is used for different numbers of subcarrier allocations. Since one of the main objectives of network coding systems is to improve the transmission rate and bandwidth efficiency, it is important to allocate different numbers of subcarriers for fixed and variable bandwidth efficiency cases in network coded OFDMA systems. Utilizing the relay in the network coded OFDMA system, the diversity order and therefore outage performance can be increased with respect to the traditional OFDMA system. Moreover, available bandwidth can be used efficiently according to the user requirements by this way. Considering the obtained results, it can be argued that the proposed techniques in this thesis will provide a different perspective and new insights into the subcarrier allocation problem of traditional and network coded OFDMA systems.