Performance Analysis And Optimization In Two-tier Femtocell Networks

In conventional cellular system, fractional frequency reuse is always used to avoid inter-cell interference, but it incurs decreased spectrum efficiency. In other words, providing reliable communication and obtaining higher spatial frequency reuse could not be achieved at the same time in cellular systems. To resolve the tradeoff and create the dual benefits, the surest way is by reducing cell radius or getting transmitter and receiver closer to each other. However, this inevitably requires deploying more infrastructures, e.g., microcells, distributed antennas, relays and et al. For the network operators, more deployment and maintenance of these infrastructures is potentially expensive. And more importantly, it is difficult for the infrastructures to cope with the increasing data traffic originating indoors. Therefore, femtocell, as a low-cost solution to guarantee high quality indoor reception and improve indoor throughput, attracts significant interests.The femtocell-aided cellular system is obviously divided into macrocell tier and femtocell tier, which is referred to two-tier femtocell networks. Because of the scarce availability of spectrum and the ease of deployment, operators will likely prefer assigning the same spectrum resource to the transmissions in both tiers. But for reason of the limited coordination between macrocell and femtocell, as well as different femtocells, concurrent transmissions will result in cross-tier and intra-tier co-channel interference, which inevitably degrades the performance of per-tier transmission and network throughput.By employing Possion point process to model the random location of femtocells, the outage performance, coverage and throughput of downlink transmissions in two-tier femtocell networks with universal frequency reuse are analyzed, which quantizes the effect of different network parameters on downlink performance and further provides theoretical basis for the optimization works. Moreover, in order to guarantee the performance of cellular users and improve network throughput, two resource management schemes in OFDMA-based system are proposed, which accelerates the femtocells delployment to provide tangible benefits. Concretely, the main contributions are:(1) The per-tier downlink outage performance is studied based on co-channel interference and pathloss model. In case of universal pathloss exponent, the tight lower and upper bounds of per-tier downlink outage probability are derived respectively by dividing the interfering femtocells into near field nodes and far field nodes. In many empirical pathloss models, the pathloss exponent in outdoor environment usually takes value as3.8-4. When the pathloss exponent of interference from femtocells equals4, since the aggregated cross-tier or intra-tier interference from femtocells follows Levy’s stable distribution, the close-form expressions on per-tier average downlink outage probability are derived.(2) Due to the effect of cross-tier interference, the femtocell close to macrocell base station could not cover the indoor region completely. As a result, the indoor users at femtocell edge handover to macrocell. But for the femtocells far from macrocell base station, the coverage extends to outdoor region, and the cellular users in the outage region are severely interrupted. Therefore, with the decision criteria based on the long-term received power, the coverage of femtocell in different scenarios is discussed. According to the modelings of expected signal and interference signal, the femtocell’s service region and coverage region can be divided into different geographic zones. The received SIR and average throughput in each geographic zone are analyzed. Furthermore, considering round robin scheduling system, the average throughput of outdoor users and indoor users are obtained respectively.(3) Employing multiple antenna transmission provides increased spatial reuse and increased robustness against near-far effects at the user terminal. However, if user terminal feedbacks channel state information for transmit beamforming strategy via delayed and limited-rate feedback channel, inter-user interference in multi-user MIMO transmissions inevitably degrades the system performance. Based on the imperfect feedback channel and multi-user transmit beamforming strategy, the probabilities of per-tier downlink successful transmissions are derived, which reflects the effect of limited-rate and delayed feedback on MIMO transmissions. Moreover, per-tier coverage and maximum femtocell contention density are discussed with QoS requirement, showing the effect of multi-user MIMO transmissions on the coverage and spatial reuse in two-tier networks.(4) In two-tier networks with universal frequency reuse, the downlink outage performance of cellular users and network throughput are degraded due to cross-tier co-channel interference. To optimize the performance in two-tier networks, two resource management schemes are proposed in OFDMA systems. The first scheme defines the limited region of cross-tier interference for cellular user, and assigns the transmissions in femtocell in the region to the subcarriers orthogonal with those in macrocell. Hence the outage probability of downlink transmissions of cellular users can meet QoS requirement. In the second scheme, the downlink transmissions in macrocell and femtocells share a fraction of available subcarriers. With the limitation that the average throughput of per-tier single user can meet QoS requirement, the average network throughput is maximized by adjusting the fraction of sharing subcarriers.In summary, the downlink performance of single-antenna and multi-antenna transmissions is studied in the context of two-tier femtocell networks with universal frequency reuse, which provides theoretical basis for performance optimizations. Furthermore, the optimization of downlink performance is studied in two-tier network, by proposing two resource management schmes, which effectively improve the outage performance of cellular user and the average network throughput.