Huawei UCN Research and Practice

Huawei UCN Research and Practice A NEW CONCEPT In recent years, DSL systems have faced many new challenges, such as crosstalk. The standardized DMTs modulation techniques are deployed more densely than ever, making radio resources allocation a severe challenge to address. Insertion of additive guard bands like cyclic prefixes led to the addition of new deal. These insertions are not enough to simply improve the capability of a single traditional DMT transceivers to counteract the impact of the crosstalk. It lies that network development requires a new way of thinking. Based on an optimization-based perspective, next generation DMTs require coordinate network nodes, frequencies and bands, and uniformly arrange network resources, and capable to provide optimal user experience. Therefore, Huawei established a new concept: User Centric Network (UCN). User Centric Network (UCN) is a concept of user-centric network construction. In traditional network construction, base stations were centered, and users were served by a certain base station. As users may be located in different places, it is a challenge to ensure stable and reliable performance for users. Interference between adjacent base stations also reduces the resource efficiency of the entire network. With the new concept of UCN, resources are coordinated, combined, and optimized in allocation, based on a user-centric philosophy so that the user experience will be enhanced. UCN is also a new user-centric concept in term of operation. In the traditional way, operators can just sell simple data packages to customers. USER BENEFITS UCN focuses on users it can provide a lot of benefits for end users. First, UCN can eliminate cell boundaries, providing noborder service experience and improving the peak and average rates. Second, UCN enables multiple cells to receive signals from terminals in a coordinated way, reducing requirements for transmit power of terminals and prolongs their standby time. Third, UCN uses flexible networks, providing customized services and tariff packages for users. UCN AND 4.5G, 5G Here we have to emphasize that UCN is a network construction concept beyond the definition of wireless technology generations. UCN and 4.5G or 5G are not simply a one-to-one relationship. UCN can be implemented phase by phase in 4.5G and 5G. For example, UCN technologies can be used in the 4.5G phase, such as distributed MIMO. Distributed MIMO uses distributed, multi-site, multiple antenna beamforming and multiuser multiplexing technologies on the RAN side to reduce interference and increase capacity. In the recent field trials, distributed MIMO proved 3- to 4- folds of cell capacity. RECENT RESEARCH ON UCN At present, the number of base stations deployed on 4G networks has reached several millions. The recent research on UCN focuses on how to apply the leading-edge UCN concept to these base stations early. We are pleased to see that the entire industry has made successful progress in UCN research. CloudRAN-based technological innovation such as distributed MIMO can ideally control intersite interference and enable extremely dense deployment of sites, without the need to upgrade terminals on live networks. 4.5G distributed MIMO has been put into trial use on live networks for advanced operators. For example, the inter-site distance of lamp pole sites on Shanghais Bund is as short as 50 m. With distributed MIMO, the data rate of cell edge users has increased from 8.2 Mbps to 15 Mbps, an improvement of 80%, and the average cell throughput has increased from 45 Mbps to 65 Mbps, an increase of 45%. Minimum Mean Square Error (MMSE) Estimation for Interference Identification We are interested in an estimate of the time-varying channel gain matrix. It is obtained by means of a statistical estimation approach that combines the measurements with (i) statistical knowledge of measurement uncertainty, and (ii) prior knowledge of spatial correlation of the interference links. We assume known positions of the transmitted and received vectors and known noise vectors from which the a priori distribution of the channel gain matrix with a mean and a covariance matrix is derived. Statistical knowledge about the channel gain vector and measurement uncertainty is exploited. Given some physical-layer measurements, an ideal linear model in which the prior distribution of the interference matrix and the uncertainty distribution is Gaussian in linear scale is derived. This model relates the measurements to the channel gain vector and therefore can be used to derive an optimal linear MMSE (LMMSE) estimator for the channel gain vector. Since interference is often assumed to have a log-normal distribution, a more realistic model in which the prior path-loss distribution is log-normal and the uncertainty distribution is Gaussian in dB scale is used. In this case, the model becomes non-linear, and therefore a closed-form linearized MMSE estimator, named linearized log MMSE (LLMMSE), is derived to estimate the channel gain vector. The results presented here show how the accuracy of interference estimation obtained from the proposed MMSE Estimator is affected by two syste m parameters, namely the Reference Signal Received Power (RSRP) uncertainty à Ã†â€™ and the channel variance à Ã‚ . The performance of the MMSE is compared to the simple least squares (LS) estimator. The simulation results in Figure 2-2 show that the proposed MMSE estimator outperforms the LS estimator. The gains are large for high noise levels or when the channel variance à Ã‚  is small. The performance in low noise situations is similar to the LS performance as in such cases the solution of the MMSE estimator converges to the one of the LS estimator. Same behaviour is observed when the channel variance is high.