Researches On Radio Resource Management In MIMO OFDM Systems

The 4^th wireless communication system proposed many novel technologies, the most important two of which are multi-input multi-output (MIMO) and orthogonal frequency division multiplexing (OFDM). MIMO can provide a huge amount of the system capacity and OFDM can well defeat the inter-symbol interference (ISI), hence the combination of MIMO and OFDM must be the strongest air interface technology for the 4^th wireless communication. However, the frequency resource will become more and more emergent for guaranteed QoS requirements of various services. The frequency efficiency always decides the commercialization of a technology and The 4^th wireless communication system is no expectation. Radio resource management (RRM) is just a technology devoting to improve the frequency efficiency and providing QoS guarantees for any service. Hence, in this dissertation, we mainly focus on researches of the radio resource management in MIMO OFDM systems.The channel capacity can give some instruments to RRM. Hence, we first discuss the ergodic capacity of MIMO OFDM systems. The space-time-frequency waterfilling scheme (STFWF) is proposed to maximize the ergodic capacity through allocating power in the space, time and frequency domains. It keeps the long-term power constant and allows the short-term power fluctuating. Compared with space-frequency waterfilling scheme, STFWF scheme has the following merits. Since the waterfilling level is calculated based on long-term channel information in the STFWF scheme, the channel statistic information is only needed which can reduce the feedback information and computational complexity. Furthermore, it can suit both flat and frequency selective fading channel. At present, RRM schemes in MIMO OFDM systems are mainly focused on PHY layer and PHY-MAC cross layer. We first make some researches on RRM in MIMO OFDM beamforming systems at PHY layer. After doing some performance analyses on several kinds of beamforming technologies, the subcarrier shared directly one is chosen to provide SDMA among multiuser. Then, the RRM mathematical module is set up in OFDMA-SDMA MIMO OFDM systems. It is different from the conventional ones by considering both space multiuser diversity and all the space subchannels for each subcarrier. Hence, it improves the frequency efficiency to the largest extent. Here, each subcarrier and its space subchannels are defined as a group of space-frequency subchannels. Unfortunately, it is a NP-hard problem from mathematical viewpoint. Hence, the suboptimal space-frequency subchannel allocation and adaptive modulation (SSFSA) is proposed which divides the RRM problem into three steps. First, subcarriers are not shared among users and the optimal space-frequency subchannel allocation outcome is obtained by mathematical deduction. Then, based on the subcarrier allocation in the first step, the subcarriers sharing scheme is proposed to share subcarriers among users with low spatial correlation at the premise that negligible co-channel interferences are introduced. Bit loading and power allocation are made at the third step for each user on its all space-frequency subchannels. Considering the computational complexity of SSFSA and feedback information reduction, a practical space-frequency subchannel allocation and adaptive modulation (PSFSA) is also proposed. This scheme defines an initial subcarrier allocation factor based on the first step subcarrier allocation outcome of SSFSA. It can reduce the computational complaexity to init subcarrier allocation according to the factor. An improved interpolation-based beamforming algorithm is also proposed to reduce the feedback information in PSFSA.Since considering both the QoS characteristics at the upper layer and the channel information at the PHY layer, cross layer RRM is an efficient method to manage resources in wireless networks. We make some researches on cross layer RRM in IEEE 802.16 MIMO OFDM systems. There are two ways to realize cross layer RRM. The first one is mapping channel information from PHY layer to upper layer and executing RRM at the upper layer. The other one is mapping QoS parameters to PHY layer and allocate resources at the PHY layer. The proposed QoS-guaranteed fairness scheme is the first type which mapping the channel fading information to the MAC layer. The subchannel allocation untility function is defined considering QoS requirements of different services, the status of scheduler, the fairness among connections and the channel fading status. The subchannel is allocated to the connection with the largest subchannel allocation unility and then bit and power allocations are made for each connection. While the optimal QoS-guaranteed resource allocation scheme belongs to the second one which maps the upper rtPS delay requirements to scheduling rate requirements. And then, the capacity-maximized resource allocation problem with scheduling and power constraints is solved by mathematical deduction. From the simulation results in Matlab, the system throughput of QoS-guaranteed fairness scheme is less than that of the optimal QoS-guaranteed resource allocation scheme, while the fairness of the former is better than the later.