Prototyping Massive MIMO Resolution for Prototyping of Massive MIMO

The wireless demand for wireless data continues to motivate developers to find new technologies to expand wireless data capacity and network capabilities. Industry experts generally believe that even if the current and planned infrastructure is fully developed, data requirements will continue to exceed existing capabilities, and the debate has shifted from “whether” to “when”. Wireless service providers are planning to upgrade their networks to 4G LTE, LTE Advanced (LTE-A), and more advanced technologies, with innovative solutions such as microcellular coverage, heterogeneous networks, carrier aggregation, and 3GPP roadmaps. However, it is clear that the capacity ramp produced by the current technology trajectory is still flatter than the demand line. Faced with this challenge, the 3GPP standard entity recently proposed the goal of “maximizing 1000 times by 2020” to meet the needs of evolutionary or revolutionary ideas.

This concept requires the base station to deploy a very large array of antennas, possibly containing hundreds or thousands of transceivers. This concept is called massive MIMO. Indeed, massive MIMO is off the current network topology and may be the key to solving the wireless data challenges we face; however, in the process of recognizing the effectiveness and/or feasibility of widespread MIMO deployments, A question worthy of attention, someone will create a prototype, just to determine if it really works? After all, creating a prototype with thousands of antennas presents several engineering challenges, plus other issues that cannot be ignored, cost and time.

Figure 1 Antenna MIMO transceiver.

MIMO background

MIMO relies on multiplexing to improve the reliability of wireless data links and effective data rates, often using multiple independent antennas to obtain multiple data streams. Multipath propagation is a huge challenge for communication systems. In practice, MIMO is used, using a variety of techniques such as space-time coding and/or spatial diversity. The 4G mobile communication standard LTE-A specifies a maximum of 8 antennas for MIMO configuration. MIMO is commonly used in the IEEE 802.11n/ac standard and the actual commercialization of these standards.

Basically, more antennas give the propagation channel a higher degree of freedom, resulting in higher performance in terms of data rate and/or link reliability. However, the overall data rate is still limited by Shannon's theory. In a network composed of multiple users, one method of increasing the overall network throughput is multi-user MIMO (MU-MIMO), in which multiple users can access the same time-frequency resource at the same time, but multiple generations are generated by multiple antennas. The "space dimension" achieves isolation.

More antennas, larger capacity, higher reliability

Increasing the size of MU-MIMO, called massive MIMO, can provide greater network capacity, higher reliability, and increase the energy efficiency of large-scale MIMO base stations by reducing the total transmit power of a cell or service area. In theory, the transmit power of each antenna can be lower than the transmit power of a single antenna served at the same data rate for a given cell or region. That is, the total power is:

PTotMM ~ PT NT

Among them, PTotMM is the total transmission power of each area, PT is the power of each antenna, and NT is the number of transmitting antennas. Among them, PTotMM is lower than PTot of single antenna system. In order to achieve the same reliability and/or throughput, a large-scale MIMO cellular base station can focus on the target user with its higher degree of freedom compared to a single antenna system. Reduce the total transmit power of the partition area. In addition, when multiple antennas are used, the probability of the correct bit transmission from the transmitter to the receiver increases because the link outage probability is ~ 1 / SNR NT NR.

Where SNR is the signal to noise ratio, NR is the number of receiving antennas, and NT is the number of transmitting antennas. Due to this relationship, when the number of antennas in the system increases, the probability of link interruption is lowered, thereby improving the reliability of the communication link. [1]

Massive MIMO antenna arrays are based on the basic concepts described here, and by theory, hundreds of times the size of antenna deployments will achieve higher efficiency than current MIMO point-to-point deployments. Specifically, with hundreds of antennas, the antenna aperture and deployment grid have finer resolutions. With beamforming, the antenna lobes can be more finely controlled to reduce the energy in the channel.

Massive MIMO systems also have their challenges. One challenge is to find a channel state information communication method from the receiver to the transmitter for precoding. Given the hundreds of antennas, inferring channel states through pilot signals is not feasible in practice. Therefore, the large-scale MIMO currently implemented can only actually use a time division duplex (TDD) system that relies on channel reciprocity. However, more research is needed to determine the feasibility of this method. In addition, some preliminary studies suggest that thermal noise in the system does not have to be too much attention for so many antennas, and the effects of jammers become a bigger problem. These and other challenges can be studied using actual waveforms after developing an effective prototype.

Figure 2. M-user N antenna massive MIMO system

Figure 3. Typical 1x1 software-defined radio architecture.

Prototyping of massive MIMO systems

Prototyping large-scale MIMO systems requires a lot of work in advance to carefully and properly design the actual operating system. Most researchers will find it even challenging to make a minimally configured MIMO transceiver system with only 2 antennas (see Figure 1). To design a large-scale MIMO prototype, first sketch the system (see Figure 2). In this exercise, the number N of antennas at the base station is 128, resulting in a 128&TImes;128 MIMO configuration. The configuration assumes that M mobile users use SISO antennas.

There are many things to consider when designing a massive MIMO system, including RF system parameters such as transmit power, adjacent channel interference, and spectrum mask. However, one of the key parameters to consider for large-scale MIMO systems is the digital data throughput of each antenna. As you can see from the figure, one of the most challenging aspects of the system is to aggregate all received samples into a common processing subsystem. Unlike simple transmit and receive communications using SISO radios, massive MIMO requires high-speed data throughput between transmit and receive components, as well as high baseband, and is orders of magnitude higher than currently deployed systems.

The data stream can be processed in a distributed manner at nodes close to the antenna, but in order to recover signals received from different users, or to effectively pre-code signals for different users, the data streams received from the various antennas must be aggregated. In a common location for optimal performance. By carefully observing throughput and data requirements, we divide the system into basic components. In this way, we can quantify the data rate in the actual construction of the prototype and balance the system design, integration, power and cost.

Baseline system parameters

A typical SISO radio is shown in Figure 3. In this figure, the RF signal is downconverted or mixed, filtered, amplified, and then converted to digital data. The order of the launch process is reversed. Massive MIMO systems contain hundreds of such basic SISO elements. To use off-the-shelf equipment to reduce costs and speed up prototyping, assume that each in-phase quadrature sample is 16 bits. The number of bits determines the dynamic range and is actually good for the prototype. Reducing the number of resolution bits can significantly reduce data throughput, especially when gathering very many channels. Although 16 bits will increase the data path and ultimately increase data throughput requirements - more bits will result in increased data path and increased data throughput requirements - however, off-the-shelf components and programming architectures do not need to be customized Able to light

Loose 16-bit samples

Next consider the sampling rate. Each analog-to-digital converter (ADC) in the receive chain must sample the downconverted waveform at a higher rate than the Nyquist channel bandwidth. In this example, LTE is used as the baseline. For normal mobile communication scenarios, each converter samples the received waveform at a sampling rate of 30.72 MS/s. In fact, the converter can oversample the signal to increase resolution, but this increases the amount of signal processing to convert the data rate to a data stream acceptable to the standard signal processing block. The data throughput is obtained using the following equation: (2 samples) (16 bits / sample or 2 bytes / sec) (sampling rate)

For the above example:

(2 samples) (2 bytes/second) (30.72) = 122.88 MB/s For the above example system, the aggregate data throughput of one channel is equal to 122.88 MB/s. To scale to a massive MIMO system, the effective rate can be calculated as follows: Total System Throughput (TST) = (throughput / channel) (number of antennas) TST = (122.88 MB / s) (128) TST = 15.7 GB /s

Thus, if all channels transmit or receive simultaneously, the central processing system's data throughput will be 15.7 GB/s. In addition, the aggregation of all of this data into the central processing system also requires the processing engine to accept this large amount of data and to process the data further to generate communication links. The above brief analysis reveals two challenges. First, very few, if any, low-cost, commercially available technologies can meet these requirements. Second, the amount of data in the prototype requires the development of alternative signal processing chain segmentation techniques, including distributed implementations and parallel implementations.

By reviewing available prototyping techniques, we present a brief study of a high-speed serial bus that can be used as a framework for building large-scale MIMO prototypes.

Table 1 summarizes some of the current commercial high-speed bus technologies. There are of course other buses, but the above table provides a guide to many of the standards that are currently in use, not proprietary bus technologies. In addition, these bus technologies have been used in many modular architectures, such as PXIe, which are basically based on the PCIe standard. One specification that should be considered is the latency. Latency is the turnaround time between transmit and receive operations. Latency is not particularly important if the prototype is for a unidirectional link. However, for true TDD massive MIMO prototypes, latency must be considered because the cycle time is shorter than the coherence time of the wireless channel, so downlink precoding is not based on outdated channel information, which is critical. The latency specifications given above are approximate. However, in general, the latency of Ethernet is not decisive and can change dramatically. On the other hand, Ethernet implementations are generally less expensive.

It should be noted that the PCIe Gen 3 implementation has just appeared on the market and the actual throughput data measurements are not available. It should also be noted that although the maximum/peak data rate is basically provided, the typical implementation of the bus is actually different due to cost, size of the IP core, and power. The typical number provided is for reference only, as very few (if any) implementations achieve the maximum rate of release.

Figure 4 shows an example of a system configuration using PXIe.

In this configuration, a total of 10 backplanes were used to implement a massive MIMO system with 128 antennas. The system uses two “primary” backplanes to aggregate data and eight backplanes to install 128 transceivers (NI 5791 RF Transceivers) that can transmit and receive on the cellular band. The data backbone uses PCI Express Gen 2 &TImes;8 to easily acquire and transmit 20MHz RF bandwidth data with appropriate segmentation.

POE ONU

In general, ONU devices can be classified according to multiple application scenarios such as SFU, HGU, SBU, MDU, and MTU.
1. SFU type ONU deployment
The advantage of this deployment method is that the network resources are relatively abundant, and it is suitable for independent households in the FTTH scenario. It can ensure that the user terminal has broadband access functions, but does not involve more complicated home gateway functions. The SFU in this environment has two common forms: it provides both an Ethernet interface and a POTS interface; it only provides an Ethernet interface. It should be noted that SFU can provide coaxial cable functions in both forms to facilitate the realization of CATV services, and it can also be used with home gateways to facilitate the provision of value-added service functions. This scenario is also applicable to enterprises that do not need to exchange TDM data.
2. HGU type ONU deployment
The HGU type ONU terminal deployment strategy is similar to the SFU type, except that the functions of the ONU and RG are integrated in hardware. Compared with SFU, it can realize more complicated control and management functions. The U-shaped interface in this deployment scenario is built into the physical device and does not provide an interface. If you need to provide an xDSLRG device, you can directly connect multiple types of interfaces to the home network, which is equivalent to a home gateway with an EPON upstream interface. Used in FTTH occasions.
3. SBU type ONU deployment
This deployment scheme is more suitable for independent enterprise users to construct the network under the FTTO application mode, and is based on the enterprise change of SFU and HGU deployment scenarios. The network in this deployment environment can support broadband access terminal functions and provide users with multiple data interfaces including El interface, Ethernet interface, and POTS interface, which can meet the needs of enterprises in data communication, voice communication, and TDM dedicated line services. Usage requirements. The u-type interface in the environment can provide enterprises with a frame structure with multiple attributes, which is more powerful.
4. MDU type ONU deployment
This deployment scheme is suitable for multi-user network construction under multi-application modes such as FTTC, FTTN, FTTCab, and FTTZ. If enterprise-level users have no demand for TDM services, this solution can also be used for EPON network deployment. This deployment solution can provide broadband data communication services for multiple users, including Ethernet/IP services, VoIP services, and CATV services and other multi-service modes, with powerful data transmission capabilities. Each of its communication ports can correspond to a network user, so in comparison, its network utilization rate is higher.
5. MTU-type ONU deployment
The deployment plan is based on the commercialization of the MDU deployment plan. It can provide multi-enterprise users with multiple interface services including Ethernet interfaces and POTS interfaces, and can meet enterprise voice, data, TDM private line services and other services demand. If combined with the slot type implementation structure, richer and more powerful business functions can be realized.

Poe Function,Outdoor Poe Onu,Optical Network Unit,POE ONU 4 Port,4GE POE ONU

Shenzhen GL-COM Technology CO.,LTD. , https://www.szglcom.com