WiMAX Wave2 Dual-Channel MIMO Measurement

WiMAX Wave 2 specifications now support the use of multiple antennas to improve system performance in both the downlink and uplink. Systems configured with multiple-input multiple-output (MIMO) configurations use spectrum more efficiently, resulting in higher data rates, than traditional single-input single-output (SISO) implementations. Characterization and troubleshooting of these advanced WiMAX systems typically requires the use of a dual-channel signal analyzer with channel estimation, a "matrix decoder," and an OFDM demodulator.

Matrix A and Matrix B Configurations

There are two implementations of multiple antennas operating on the downlink transmit side of a WiMAX Wave 2 system: Space Time Coding (STC)—Matrix A and MIMO—Matrix B. Figure 1 shows a typical downlink configuration for 2x1 STC and 2x2 MIMO.

In the Matrix A (STC) implementation, the channel can be modeled as two paths connecting the two transmit antennas of the base station (BS) to one receive antenna at the mobile station (MS). Each signal path can be represented by a unique channel coefficient "hx". Each coefficient represents the (assumed linear) set of all paths between each transmit-receive antenna pair, and may also include inter-channel crosstalk generated in the transmitter and the countless multipath signals that appear in the wireless channel. In addition, the signal reception quality can be improved by using each antenna to transmit different coded versions of the same signal at the same frequency at different times. This technique is called spatial diversity, and it is implemented by the Matrix A configuration.

In contrast, the Matrix B (MIMO) system uses each antenna to transmit different data streams on the same frequency channel at the same time to achieve higher data rates and spectral efficiency. For the Matrix B configuration shown in Figure 1, the received signal measured in a noise-free system is:

(1)

(2)

Assuming the four channel coefficients are known, a matrix B receiver can use the following simplified method to discern and recover the transmitted waveform.

(3)

(4)

(5)

These equations can also be expressed in matrix form:

(6)

The function of the matrix decoder is to perform the inversion of the channel matrix [H] and related mathematical operations to recover the originally transmitted data stream and pass this information to the demodulator. Note that matrix decoding is independent of demodulation and is performed prior to demodulation.

When the channel coefficients are correlated, actual WiMAX receivers can use their inherent decomposition or MMSE techniques [Reference 1] to perform true data recovery. As mentioned above, data recovery requires the knowledge of the channel coefficients, which can be measured by a receiver or a dual-channel signal analyzer using the unique pilot structure contained in the WiMAX OFDM waveform [Reference 2]. It is important to note that accurate matrix decoding depends on the degree of independence of the channel coefficients and is further affected by the amount of noise in the channel. When the channel matrix becomes "ill-conditioned" and is difficult to invert accurately, the correlated channel coefficients and/or noise will cause system performance to degrade.

In the uplink, MIMO is implemented by coordinated, synchronized transmissions between two independent mobile stations (handsets) operating on the same frequency channel. This technique, called uplink cooperative spatial multiplexing (UL-CSM), uses two or more receive antennas at the base station and one antenna at each mobile station to implement 2x2 MIMO [Reference 2]. In this configuration, MIMO implementation is limited to the uplink. DL-MIMO requires two antennas and receiver channels at each mobile station.

Channel Estimation, Matrix Decoding, and Demodulation

Signal analysis and troubleshooting of the Matrix A and Matrix B waveforms can be performed using a single-input or multi-input vector signal analyzer (VSA). Many basic measurements, such as STC or inter-channel crosstalk and timing in a MIMO transmitter, can be performed by connecting a single-input analyzer directly to the selected transmitter output port [Reference 3]. This single-input approach is useful when the transmitted signals have good isolation and do not require the use of a matrix decoder to demodulate the waveform. Some test procedures, such as the Radio Frequency Conformance Test (RCT) defined in the WiMAX Wave 2 profile, specify single-channel measurements of transmitter signal quality in the presence of crosstalk and without the use of a matrix decoder. Unfortunately, this basic measurement is of little use in analyzing the root cause of many signal errors during system optimization and troubleshooting. In such cases, finding the root cause of the error often requires comparing measurements taken with and without the matrix decoder. In a Matrix A system, a single-channel VSA can be used to test both with and without the decoder. In Matrix B and UL-CSM systems, a dual-channel VSA is typically required to fully analyze these increasingly complex waveforms.

Figure 2 shows a typical measurement flow for a dual-channel VS with WiMAX MIMO measurement capability (such as the Agilent 89600 Series analyzer with Option B7Y). In the Matrix B configuration, MIMO signal analysis begins with the estimation of complex channel coefficients, which are obtained by measuring a large number of known pilot subcarriers received at the two input signals, as shown in the figure for Rx0 and Rx1. These four channel coefficients are displayed as a function of subcarrier frequency and are a very useful analysis tool when optimizing and troubleshooting MIMO systems. The estimated channel coefficients are primarily used by the matrix decoder to recover the two independent data streams in the 2x2 MIMO signal. The matrix decoder is responsible for canceling channel effects rather than performing data demodulation. As shown in Figure 2, the matrix B data streams are recovered and then transmitted to the OFDM demodulator for further signal analysis.