Emulation methodology of multiple clusters channels for evaluating wireless communication devices over-the-air (OTA) performance is investigated. This methodology has been used along with the implementation of the SIMO LTE standard. It consists of evaluating effective diversity gain (EDG) level of SIMO LTE-OFDM system for different channel models according to the received power by establishing an active link between the transmitter and the receiver. The measurement process is set up in a Reverberation Chamber (RC). The obtained results are compared to the reference case of single input-single output (SISO) in order to evaluate the real improvement attained by the implemented system.
In recent research works, reverberation chamber (RC) is considered as a useful tool to emulate rich multipath environments [
Active measurement methods are often based on the use of a channel emulator associated with a real-time transmission system, to test the operational terminals. In this paper, the aim is to suggest an experimental platform using a small size reverberation chamber (reverberant cell) to study the feasibility of emulating multipath channel while maintaining a Rayleigh fading (in order to be able to compare different receivers in reference environments with the same distribution). On this platform, a multicluster emulation method which complies with channels defined by 3GPP models is implemented, using only one vector signal generator. This emulation must be accompanied by a strict control of delay spread, to generate realistic channels [
The presented approach aims to develop a flexible OTA methodology for quantification and implementation of digital multiantenna transmission systems inside a small size reverberant cell. On this test bed, the measurements are not carried out in real time and are not dedicated to performance evaluation in terms of throughput. However, it allows (through the use of an RF digitizer and baseband processing in MATLAB) to study in detail the influence of several transmission chain parameters, as antenna aspects (in MIMO context: coupling, correlation coefficient
In this paper, this method is applied to test the 3GPP LTE standard, by implementing an LTE-OFDM frame and using diversity at the receiver side. The frame is generated based on the 3GPP standard [
LTE also uses adaptive modulation and coding to get better data throughput. The modulation schemes supported for payload in the uplink and downlink are QPSK, 16QAM, and 64QAM [
As shown in Figure
LTE Frame [
After having introduced the LTE requirements, it is necessary to focus on parameters for evaluating the performance of a system under test. In this paper, a SIMO configuration is considered, with two synchronized receivers.
The fundamental parameters usually used to estimate the diversity performance are the correlation coefficient [
The different parts discussed in this paper are as follows. Section
The LTE signal described previously is implemented on the measurement test bed in Figure
Measurement test bed.
The measurement test bed is based on the Aeroflex PXI 3000 series architecture, with two PXI chassis integrating a control PC for generating frames on transmission and for processing received data. The transmit part includes one RF wideband signal generator (76 MHz–6 GHz), which can provide a level of RF power from −120 dBm to +5 dBm over a modulation bandwidth of 33 MHz. The receiver integrates two digitizers, which provide conversion of RF signal to baseband digital IQ symbols [
At the transmitter side, a frame based on LTE specifications is generated. The duplex mode used is TDD, and a bandwidth of 5 MHz has been chosen, with 64QAM modulation scheme, over a carrier frequency of 2.35 GHz.
The characteristics of the LTE-OFDM frame and measurement system to be used has been explained previously. This section will focus on establishing a method for the emulation of 3GPP channel models with a specific delay spread, which requires a control of the delay spread inside the RC.
The 3GPP urban microcell and urban macrocell channel models in [
Urban micro- and macrocell channel model taps.
It can be seen that the urban macrocell channel presents a high delay spread compared to the urban microcell.
The taps delay and the power magnitude are listed in details in Table
Parameters of the urban microcell and urban macrocell scenarios.
Scenario | Urban Macro | Urban micro | ||
Relative power (dB) | Delay ( | Relative power (dB) | Delay ( | |
Taps Model | 0 | 0 | 0 | 0 |
−2.22 | 0.36 | −1.27 | 0.28 | |
−1.72 | 0.25 | −2.72 | 0.20 | |
−5.72 | 1.04 | −4.30 | 0.66 | |
−9.05 | 2.7 | −6.01 | 0.81 | |
−12.50 | 4.59 | −8.43 | 0.92 |
Some special consideration should be taken into account when implementing channel models in an RC. Indeed these models introduce intracluster delay spread. Then, for each tap an RMS delay spread
In order to obtain the desired
The RC used in this work is the SMART 1000 Mini Reverb-cell [
Reverberation chamber overview.
The delay spread
The measurements shown in Figure
Measured PDPs for unloaded and loaded RC.
For the unloaded chamber case, the measured RMS delay spread is equal to 256.4 ns. When the chamber is loaded with a piece of material, which is solid, pyramidal shaped, and carbon loaded (urethane foam absorber [
In order to check the fading distribution, a power measurement is made at the central frequency to plot the normalized CDF. Figure
Cumulative distribution function measured in the loaded and unloaded RC.
In this part, the control of the
The method to emulate the channel model consists of convolving the base band signal to be transmitted with the urban macro-cell or urban micro-cell channel model tap delay line generated using a MATLAB program (see Figure
Channel emulation overview.
In order to verify the proper functioning of this method, a channel sounding based on a sliding correlation [
Figures
Power delay profiles measured in RC for urban micro-cell (a) and urban macro-cell (b) channel models.
The different results presented in this section highlight the possibility of controlling delay spread for each cluster, and emulating 3GPP urban micro- and macro-cell channel models. This is obtained by combining a digital preprocessing and a RC to manage the
In order to realize these measurements, two boards representing a compact terminal are used: one reference board with one triband antenna, and diversity board with two triband antennas (see Figure
Diversity system under test characteristics.
Frequency (GHz) | Return losses S11 (dB) | Coupling coefficient S21 (dB) | Correlation coefficient (in isotropic environment) | Total efficiency (%) |
---|---|---|---|---|
2.35 | −12 | −17 | <0.05 | 92.7 |
Reference antenna (a) and diversity antennas (b) [
By using the LTE system implemented in Section
The CDF curves plotted in Figure cumulative distribution calculated from the measured power at one frequency in the signal bandwidth, the CDFs from the averaged power of received signal, demodulated after equalization.
SISO cumulative distribution function measured in an RC for one cluster, urban micro- ,and macrocell channel model cases.
For this comparison, each measured power of these results is normalized to its own average power.
As it is known OFDM signal gives better performance than using mono-carrier signal. It is an efficient way to deal against intersymbol interference, because such interference affects only a small percentage of the subcarriers. These interferences cause a frequency selectivity that can be well observed in Figure
These results can be confirmed by the signal-to-Noise ratio measurement, presented below. The estimation of the measured mean SNR presented in Table
Measured mean SNR value for SISO configuration for a 64-QAM modulation scheme.
Monocluster | Urban microcell | Urban macrocell | |
---|---|---|---|
Mean SNR (dB) | 25.7 | 7.9 | 6.4 |
The SNR is the inverse of the EVM taken in dB, which is given by
In this section, the tests of the diversity systems performance are under interest. Diversity combining techniques such as Maximum Ratio Combining (MRC) have been implemented.
Multiple antennas are expressed as a
The example shown in Figure
The calculation of the fading correlation coefficient for the different channel models generated in the RC is presented in Table
Correlation coefficient.
Scenario | |||
One cluster | Urban micro | Urban macro | |
Fading correlation coefficient | 0.15 | 0.21 | 0.32 |
It can be observed that for each channel model, the fading correlation remains low, which will lead to optimal diversity results.
Regarding the effective diversity gain (EDG), it is typically calculated for a particular frequency (passive measurements). EDG is the figure-of-merit (FOM) typically used to evaluate the efficiency of the diversity antenna system [
The CDFs of the received power (for the diversity and reference systems) are depicted in Figure
SIMO cumulative distribution function measured in an RC for one cluster (a), urban microcell (b), and urban macrocell (c) channel model cases.
EDG for MRC technique at (1%) for one cluster and urban micro- and macrocell channel model cases.
In conclusion, as we can observe in Figure
In this paper, we achieve to emulate the 3GPP channel models in a small reverberation cell. The performance of an LTE SISO/SIMO-OFDM system has been studied through active measurements for different channel models with Rayleigh fading distribution. The work developed in this paper permits to evaluate capacity of a wireless system in controlled environments and gives the opportunity to test different transmission chain parameters (base band processing, modulation scheme, antennas, etc.) in the same environment and compare their performance. This approach could be of particular interest for OTA characterization of multiantenna terminals in reverberation chambers, because it helps to understand the influence of the channel characteristics under which measurements have been performed.
Future works will consist in making experiments with specific MIMO schemes, based on LTE specifications. Regarding the control of the delay spread, an analytical method will be developed to estimate it in function of the amount of absorbing material inside the RC. This method will lead to avoid channel sounding measurements, currently needed to achieve the desired delay spread.