Cloud Multiple-Input Multiple-Output Technology (C-Mimo) for 5G Communication

Compared with 4G communication, 5G wireless communication should provide thousand-fold data volume. The cloud radio access network (C-RAN) has the potential for meeting this challenge by a centralized control and processing to all the base stations through optical fibers. However, in current heterogeneous network (HetNet), small cells are assumed to be flexibly deployed within macro cell, these small-cell base stations may not be connected by the optical fiber. Mobile base stations do not have optical fiber connections as well. Moreover, the deployment of the infrastructure for optical fibers are time-consuming, especially in 5G communications, a denser network is expected and thus the number of fiber connections increases. We propose a hierarchal C-RAN, where only macro base stations directly are connected to the cloud through the optical fibers. The base stations of small cells and mobile base stations are connected to the macro base stations in close vicinity through wireless backhaul (e.g., wireless communication using millimeter wave), forming a cloud multiple-input multiple-output (C-MIMO) system. The mobile devices in this C-MIMO system simply broadcast its wireless signals. The pico/femto/mobile base stations receive the signals and relay them to the macro base stations via wireless backhaul. Then, the optical backhaul is used to send the collected signals to the cloud. The C-MIMO system is expected to have the following contributions: 1. Allow more base stations to indirectly connect to C-RAN, leading to better performance of C-RAN. 2. Reduce the amount of optical backhaul whose deployment is time-consuming, and thus speeding the realization of 5G C-RAN. 3. Upgrade 4G HetNet and can be used for 5G C-RAN, and thus enhancing the performance of the current communication system before the real 5G communication is successfully deployed. To overcome the limited capacity of the wireless backhaul, we apply the open-loop communication in C-MIMO system, where feedback information is dispensable. We exploit the diversity of signals propagating through different relay paths toward the cloud. We turn the interference in the wireless backhaul and broadcasting into the so called path diversity so as to provide reliable open-loop communication. During the three-year period, we conduct extensive theoretical analyses to establish the framework of the C-MIMO technology, and point out its design criteria. Practicability is the first priority of our project. Hence, we not only design a practical transmission technique, but also aim at efficient C-MIMO detectors. The C-MIMO detection should be able to jointly recover the data and identify the null entries of the channel matrix. Both the uplink and downlink C-MIMO detectors are tailored, where various detection algorithms like sphere decoder, Markov chain Monte Carlo, random set theory are studied. The adaptive MIMO detection, which automatically switches its detection algorithm on-the-fly depends on the channel condition, is integrated in the C-MIMO detector to reduces average complexity. The iterative receiver may also be considered. Through architecture design and its FPGA implementation, we jointly design detection algorithm and the architecture. The FPGA evaluation allows us to estimate the hardware cost of the designed C-MIMO detectors.

Project ID:PB10408-6878
External Project ID:MOST104-2218-E182-004