Comparison Between OFDM and STDCC Mobile Channel Sounders at 3.5 GHz

— In this paper, we present a comparison between OFDM and STDCC sounding techniques used to wide band channel characterization in suburban environments at 3.5 GHz band. Measurements of the received signal’s power delay profile and power level for each of the sounding techniques were obtained. These experimental results are presented and compared. For this comparison, important wide band characterization information are obtained and the performances of each one of used techniques are observed.


I. INTRODUCTION
In last years the demand for broadband services has grown significantly.The people are enjoying wireless internet access for telephony, radio and television services when they are in fixed, mobile or nomadic conditions.The rapid growth of wireless internet causes a demand for high-speed access to the World Wide Web.WiMAX (Worldwide Interoperability for Microwave Access) appears as a potential technology to attend this demand.WiMAX has potential success in its line-of-sight (LOS) and non line-of-sight (NLOS) conditions which operating below 11 GHz frequency.In Brazil, WiMAX operate on licensed frequency of 3.5 GHz, where curiously few studies are found, which makes a motivator for the work developed here.
As WiMAX, other technologies using the 3.5 GHz band has emerged in last years, therefore, the study of performance of the electromagnetic waves propagation in this band becomes essential.Furthermore, to implement designs and planning of WiMAX communication systems, the accurate propagation characteristics of the environment should be known.Channel characterization is an essential part of performance prediction for wireless communications systems.Two parameters are important in propagation prediction: the large-scale path loss and small-scale fading statistics.
The large-scale effects determine a power level averaged over an area of tens or hundreds of meters, therefore, the path-loss information is vital for the determination of coverage of a base-station (BS) placement and in optimizing it.
The small-scale parameters usually provide statistical information on local field variations and this,  The sounding techniques compared in this work have been widely used for mobile radio channel characterization through the determination of time dispersion parameters like: mean delay, delay spread and coherence bandwidth; and the frequency dispersion parameters like: mean Doppler shift, Doppler spread and coherence time [7,8,10,12].Although these techniques are very used, their results had not yet been compared for the same measurement environment.The STDCC is a time domain sounding technique while the OFDM is a frequency domain sounding technique.This difference has motivated the comparison proposed.
The organization of this paper is as follows.The two sounding techniques are summarized in Section II.In Section III, we present the morphology and topology of the environment as well as the characteristics of the equipment.In Sections IV and V, we present the transmitted and received signal in each sounding technique respectively.Section VI shows the channel characterization of the each sounding and its respective power delay profiles.Conclusions are drawn in Section VII.

A. STDCC Channel Sounder
The STDCC (Sweep Time Delay Cross Correlation) is a pulse compression technique based on the channel sounding using the property of the autocorrelation function of white noise.It is well known that if white noise n(t) is applied to the input of a linear system, the output r(t) is as follows: And if ‫ݎ‬ሺ‫ݐ‬ሻ crosscorrelated with a delayed replica of the input ݊ሺ‫ݐ‬ − ߬ሻ then the resulting crosscorrelation coefficient is proportional to the impulse response of the system ℎሺ‫ݐ‬ሻ evaluated at the time delay ߬. ‫ܧ‬ሾ‫ݎ‬ሺ‫ݐ‬ሻ݊ * ሺ‫ݐ‬ − ߬ሻሿ = ‫ܧ‬ න ℎሺߞሻ݊ሺ‫ݐ‬ − ߞሻ݊ * ሺ‫ݐ‬ − ߬ሻ݀ߞ൨ (2) As we see the impulse response of a linear system can be evaluated using white noise, and method of correlation processing.However, in practice, it is unrealistic to generate white noise, therefore, is used a m-sequence, alternatively known as pseudo-random binary sequences (PRBS) or pseudonoise (PN) sequences, because possess excellent periodic autocorrelation properties [3], as illustrated in Figure 1, which shows a normalized autocorrelation function.
STDCC pulse compression is done by correlating the received PN-sequence with an identical PNsequence clocked at a slightly lower rate.The difference in chip rate is called slip rate, which is defined as: The difference in chip rate results in different time bases between the transmitted and the received sequences.The slower sequence (received sequence) will be aligned with the transmitted sequence again after a duration: The time domain representation, k gives us the scale factor, and it is defined as: the scale factor k is a measure of the bandwidth compression of the sounder technique adopted.The STDCC technique originally proposed by Cox [4] is employed using a PN-sequence of m=511 bits and f Tx =10MHz, generated using MATLAB software, and converting into an electrical signal with IQ-PRODUCER software.The signal is transmitted using equipment from Anritsu MG3700.A problem of the PN-sequence is the dynamic range and the side lobes, to improve the performance of it was decided to use the power amplifier and filtering in the sequence before passing it.
The characteristics of the PN-sequence used are shown in Table I.The rate difference between the transmitter and receiver is 10KHz, therefore, the scale factor k is 1000.

B. OFDM-Based Channel Sounder
Orthogonal frequency division multiplexing (OFDM) also known as multicarrier modulation, incorporates a large number of orthogonally selected subcarriers to transmit a high-data rate stream in parallel in the frequency domain, as shown in Figure 2. The multipath effect introduces inter-symbol interference (ISI) that is significantly reduced in OFDM due the parallel data transmission through multiple carriers.Spectral efficiency and multipath immunity are two major features of the OFDM technique.The Figure 3 shows a transmitter and receiver OFDM.We can observe in Figure 3.a that serial QAM symbol is first converted to N s parallel subcarriers that are buffering and mapping in N s complex subsymbols d i which determine the points of the constellation of each subcarrier according to each type of modulation used, then is applied a IFFT [5].
The mathematical expression for s(t) is defined as the following: where ܶ is the OFDM symbol time.The subcarrier spacing is defined as 1/ܶ , which makes them orthogonal over one symbol period.A Guard Interval is inserting to avoid Inter-Symbol Interference (ISI).Two techniques are commonly used to represent this guard time: Cyclic Prefix or Zero-Padding.
With the addition of a Cyclic Prefix an OFDM system offers inherent robustness to multipath dispersion with a low-complexity RX.The length of the Cyclic Prefix determines the amount of multipath energy captured.Multipath energy not captured during the Cyclic Prefix window results in ICI.Therefore, the Cyclic Prefix length needs to be chosen to minimize the impact due to ICI and maximize the collected multipath energy, while keeping the overhead due to the Cyclic Prefix small.Most conventional wireless OFDM-based systems use a Cyclic Prefix to provide robustness against multipath.However, the same multipath robustness can be obtained by using a zero-padded (ZP) prefix instead of the Cyclic Prefix.
The only modification that is required at the RX is to collect additional samples corresponding to the length of the prefix and to use an overlap-and-add method to obtain the circular convolution property [10].The OFDM-based sounder is employed using the MG3700A equipment of the Anritsu.The OFDM signal was generated using MATLAB software, and converting into an electrical signal with IQ-PRODUCER software.Figure 4 shows a comparison between OFDM, filtered PN-Sequence and unfiltered PN-Sequence signals.

III. ENVIROMENT AND MEASUREMENT EQUIPMENT
A measurement campaign was performed in a residential area in Rio de Janeiro, Brazil.The environment can be characterized as suburban.The transmitter is located at a fixed position at a height of 33 m and frequency of 3.5GHz.The building chosen was Gama Filho University main building.A measurement run was performed along four routes, as shown in Figure 5. Transmitting power was -10 dBm, the antenna gain was 15 dBi and power amplifier gain was 37.8 dB.IQ-PRODUCER software.The electrical signal is amplified and after that sent to the antenna to be radiated in free space.For STDCC Sounder, it was used a Remez filter for reducing the sidelobes [5] and thereby improving the dynamic range, thus achieving a better resolution of the sounder, it can be seen in Figure 4.
The reception sounder consists of a signal analyzer Anritsu MS2781B, the antenna gain of 0 dBi and LNA gain of 15 dB.The received signal is digitalized and recorded for later post-processing in Matlab.Figure 6 shows the transmitter and receiver scheme.
The characteristics of the sounders were adjusted to similar performance.OFDM Sounder used 512 subcarriers, while STDCC sounder used a PN-sequence of 511 bits.Thus, the same characteristics of sounders were kept the same for comparison.

IV. EXPERIMENTAL RESULTS
The channel characterization measurements discussed in this section were performed using a radio channel sounding system that was described previously in Section III.This system operates at 3.5 GHz, has a maximum multipath resolution of 51.1 µs and a minimum multipath resolution of 0.1 µs for both sounders.
The experiment was conducted using a single transmission antenna on the last floor (elevation about 33 m) of Gama Filho University building and a mobile receiver traveling along of four routes at mean speed of 40 km/h.The data received were processed in MATLAB, the data acquisition was performed with 200µs capture time at a rate of 50MHz, with a clock synchronized with GPS, thus generating a second GPS measurement and other data received.Figure 7 and 8 shows receiver signal by STDCC and OFDM sounders respectively.Power delay profiles were obtained in two different ways according to the selected sounder.For the STDCC sounder the profile is obtained by the correlation between transmitted and received PNsequence.However, when using OFDM sounder, a random signal is transmitted using the technique of cyclic prefix as guard time.Figures 9-14 show a comparison between examples of power delay profiles with its respective multipath valid for each type of sounder and for each route.
When working with the measured power delay profiles to determine the dispersion parameters channel, we must ensure stationary at least in small paths.
Power delay profiles are obtained after using the CFAR (Constant False Alarm) technique [6].This technique allows distinguishing between real components and multipath noise components of the multipath.CFAR technique determines the noise by median and standard deviation of the power delay profile to provide varying levels of noise.The noise floor is determined by the difference between maximum peak and average plus standard deviation.
After determining the noise floor we have to evaluate: if a delay exceeds the noise floor, if the delays before and after also exceed it.However, for the power profile is valid at least one of the neighbors must satisfy the above condition.Figure 9 shows a profile measured on Route 1.The Route 1 is a route LOS typically urban.The values in Table II are consistent with obtained by Ron [7] and Matos [8].
In figures 9.a, 10.a, 11.a, 12.a, 13.a and 14.a are shown three power delay profiles for each one route.The color indicates the type of delay profile: the green, the delay profile captured; the red, the profile before and the black the delay profile a posteriori.The technical of cleaner profile CFAR [6] that it obliges us have information of the subsequent and previous delay profile.Figures 9.b

DISPERSION OF THE CHANNEL PARAMETERS TO ROUTE 2
The Route 3 begins NLOS at Amaro Cavalcanti Avenue passing below the Yellow Line and entering LOS with the transmitter on the Manoel Vitorino Street.Figures 12 and 13 present the power delay profile with its respective values of mean spread and RMS delay spread presented by Table IV.

DISPERSION OF THE CHANNEL PARAMETERS TO ROUTE 3
Route 4 is NLOS.Figure 14 illustrate one of the power delay profiles over the route, and Table V shows the values of mean spread and RMS delay spread in this route.
Only the STDCC sounder has shown a satisfactory performance in LOS cases.The OFDM sounder did not perform at all in this situation as can be seen in the tables II to V.

Fig. 3 .
Fig. 3. OFDM Transceiver Architecture.The receiver performs the reverse operation to recover the signal, as shown in Figure 3.b.The OFDM signal bandwidth is a function of the number of subcarriers as well as the subcarrier bandwidth.

Fig. 5 .
Fig. 5. Satellital view of the measurement environment and routes.STDCC and OFDM sounders schemes use the same equipments, only change the signal generation and post-processing software.For STDCC sounding technique, a PN-Sequence was generated and for OFDM sounding technique, it was generated a random signal according section II.B.Both signals were generated in MATLAB software and after that they were converted into an electrical signal with

Figure 7
Figure7shows a pseudorandom characteristic consistent with that expected for STDCC Sounder and Figure8shows the OFDM signal, including the gap.It was inserted to facilitate synchronization of OFDM sounder.

TABLE I .
SPECIFICATIONS PN SEQUENCE TX AND RX

TABLE II .
DISPERSION OF THE CHANNEL PARAMETERS TO ROUTE 1Route 2 is NLOS.The profiles are illustrated in Figure11, the values of mean spread and RMS delay spread are shown in TableIII.We note that the OFDM sounder did not show any results in the LOS case.

TABLE IV .
Works jointly with Inmetro in projects relative to Cognitive Radio and modeling of the wideband wireless channel.Consultant on planning, optimizing of wireless networks for broadband access, coordination on measurements and official reports of non-ionizing radiation for different Brazilian network operators.