Experimental evaluation of theperformance of 2×2 MIMO-OFDM for vehicle-to-infrastructure communications
- Okechukwu J. Onubogu^{1}Email author,
- Karla Ziri-Castro^{1},
- Dhammika Jayalath^{1} and
- Hajime Suzuki^{2}
https://doi.org/10.1186/s13638-015-0411-5
© Onubogu et al.; licensee Springer. 2015
Received: 1 December 2014
Accepted: 14 May 2015
Published: 25 June 2015
Abstract
In this paper, a novel 2×2 multiple-input multiple-output orthogonal frequency division multiplexing (MIMO-OFDM) testbed based on an Analog Devices AD9361 highly integrated radio frequency (RF) agile transceiver was specifically implemented for the purpose of estimating and analyzing MIMO-OFDM channel capacity in vehicle-to-infrastructure (V2I) environments using the 920 MHz industrial, scientific, and medical (ISM) band. We implemented two-dimensional discrete cosine transform-based filtering to reduce the channel estimation errors and show its effectiveness on our measurement results. We have also analyzed the effects of channel estimation error on the MIMO channel capacity by simulation. Three different scenarios of subcarrier spacing were investigated which correspond to IEEE 802.11p, Long-Term Evolution (LTE), and Digital Video Broadcasting Terrestrial (DVB-T)(2k) standards. An extensive MIMO-OFDM V2I channel measurement campaign was performed in a suburban environment. Analysis of the measured MIMO channel capacity results as a function of the transmitter-to-receiver (TX-RX) separation distance up to 250 m shows that the variance of the MIMO channel capacity is larger for the near-range line-of-sight (LOS) scenarios than for the long-range non-LOS cases, using a fixed receiver signal-to-noise ratio (SNR) criterion. We observed that the largest capacity values were achieved at LOS propagation despite the common assumption of a degenerated MIMO channel in LOS. We consider that this is due to the large angular spacing between MIMO subchannels which occurs when the receiver vehicle rooftop antennas pass by the fixed transmitter antennas at close range, causing MIMO subchannels to be orthogonal. In addition, analysis on the effects of different subcarrier spacings on MIMO-OFDM channel capacity showed negligible differences in mean channel capacity for the subcarrier spacing range investigated. Measured channels described in this paper are available on request.
Keywords
1 Review
1.1 Introduction
Multiple-input multiple-output (MIMO) systems have attracted considerable attention due to the increasing requirements of high capacity, spectral efficiency, and reliability in wireless communications. For example, MIMO systems have been adopted in the Long-Term Evolution (LTE) system, and it is expected that the upcoming developments in IEEE 802.11p and Digital Video Broadcasting Terrestrial (DVB-T) wireless standards will include the use of MIMO. It has been shown [1] that MIMO, when deployed in a rich scattering environment, is capable of achieving high spectral efficiency, capacity, and reliability by exploiting the increased spatial degrees of freedom. MIMO is often combined with the orthogonal frequency division multiplexing (OFDM) in modern wireless standards in order to achieve higher data rates and performance improvements in a multipath fading environment without increasing the required bandwidth or transmission power.
Efficient vehicular communication is a key in the development of intelligent transport systems (ITS) and requires the exchange of messages between two vehicles (vehicle-to-vehicle or V2V communications) or between a vehicle and a roadside unit (vehicle-to-infrastructure or V2I communications). Basically, there are two kinds of vehicular applications: those dedicated to providing safety services and others for non-safety applications [2]. For safety purposes, the use of licensed band at 5.9 GHz has been considered to avoid the problem of interference typically faced in the use of industrial, scientific, and medical (ISM) radio bands. Non-safety applications use the ISM band for the purpose of infotainment (e.g., high data rate Internet access for video streaming) where the availability of the service is expected to be opportunistic. The 920 MHz ISM band in Australia occupies 918–926 MHz. Among the ISM bands, the 920 MHz band gives an optimal trade-off of robustness against slow fading, achieving a longer range in cluttered environments and having a sufficient bandwidth for the high-data-rate Internet access. This paper focuses on the non-safety V2I applications at the 920 MHz ISM band which promises to provide infotainment applications, mobile internet services, and social network applications which are widely used in people’s daily activities in vehicles. The successful deployment of commercial MIMO systems will require a solid understanding of the channel characteristics in which it will operate. In order to assess the performance of new wireless communication systems using MIMO antennas, it is desirable to evaluate them in realistic measurement scenarios. Consequently, numerous MIMO channel measurement campaigns have been carried out in vehicular environments [3–7]. However, only a few research publications have considered MIMO V2V channels [8–12], and even fewer theoretically based research works have investigated MIMO V2I channels [13, 14]. A number of single-input single-output (SISO) antenna V2V and V2I channel measurement campaigns have been conducted [15, 16]. However, to the best of our knowledge, we are not aware of any MIMO-OFDM measurement results for V2I communications published in the scientific literature to date. In this paper, we focus on presenting the results of an experimental investigation of 2×2 MIMO-OFDM channel measurements performed in a real V2I driving scenario under both line-of-sight (LOS) and non-LOS (NLOS) conditions at the 920 MHz ISM band in a suburban environment. A channel sounding system based on a software-defined radio (SDR) platform was implemented and used to perform an extensive measurement campaign in a suburban environment. In comparison to the use of the conventional heavy and expensive radio frequency (RF) test equipment such as signal generators, vector network analyzers, and spectrum analyzers, SDR provides a flexible, inexpensive, and cost-effective measurement setup implemented in software that enables researchers to use and control the radio signal through software tools such as MATLAB.
The rapid development of MIMO systems has been based on the assumption that independent and identically distributed (i.i.d) or correlated Rayleigh fading with NLOS components is available and a high number of multipath components are created by the surrounding environment [17–19]. This, however, is not valid in all cases, and it is violated due to the existence of a LOS component that is stronger than other components. Hence, the channel can be more effectively modeled using the Ricean distribution. Conventionally, the presence of a LOS component is thought to limit the benefits of MIMO systems because of the rank deficiency of the channel matrix [20, 21]; however, a number of investigations [13, 14, 22–26] have shown that using antennas positioned or spaced in such a way that the LOS MIMO subchannels are orthogonal results in a full-rank MIMO channel matrix and therefore high-capacity channels. The common idea behind these approaches is to place the antenna elements sufficiently far apart so that the spatial LOS MIMO subchannels become orthogonal with a phase difference of π/2. The optimal spacings can be worked out via simple geometrical tools, while the channel matrix becomes full rank and delivers equal eigenvalues. This is known as an optimized LOS MIMO system [24]. We can determine the required inter-element spacings to achieve the maximum 2×2 MIMO capacity. The formula is a function of the inter-element distance, the transmitter-to-receiver (TX-RX) separation distance, the orientation of the arrays, and the carrier frequency.
This paper validates the theoretical maximum LOS MIMO capacity criteria in [14, 25, 27, 28] by presenting a measurement-based analysis of mean MIMO capacity (mean over the channel bandwidth and the time of 50 ms) as a function of TX-RX separation distance. The technique is based on the achievement of spatial multiplexing in scenarios by creating an artificial multipath not caused by physical objects but rather by deliberate antenna placement or separation of the antenna elements in such a way that a deterministic and constant orthogonal multipath is created at a specific TX-RX separation distance called D _{opt}. This paper also analyzes the MIMO channel capacity for three different subcarrier spacings: large subcarrier spacing (LSS), medium subcarrier spacing (MSS), and small subcarrier spacing (SSS). These subcarrier spacings approximately correspond to IEEE 802.11p Wireless Access in Vehicular Environment, LTE, and the 2k version of the DVB-T standard. It is important to note that this paper analyzes the capacity of the MIMO channels with three different values of subcarrier spacing and not the capacity of the whole system. In this analysis, the MIMO channels are estimated by the least square (LS) channel estimation method with known channel training symbols [29]. The channel estimation error due to lower signal-to-noise ratio (SNR) at a longer TX-RX separation distance is substantially reduced by applying two-dimensional discrete cosine transform (2D DCT)-based filtering to take advantage of the time and frequency coherence of the channel [30–32].
The remainder of the paper is organized as follows. Section 1.2 describes the channel model, MIMO channel capacity, the derivation for maximum LOS MIMO channel capacity criteria, and the LS channel estimation. Section 1.3 presents the MIMO-OFDM V2I measurement equipment, measurement environment, and parameters. In Section 1.4, we present the analysis of measurement results. Finally, Section 2 summarizes the paper and adds concluding remarks.
1.2 Channel model and capacity
1.2.1 1.2.1 LOS MIMO channel model
1.2.2 1.2.2 MIMO channel capacity
where γ _{ m } ( H _{ k }) is the mth eigenvalue of \(\mathbf {H}_{k}\mathbf {H}_{k}^{H}\). The maximum capacity is achieved when the channel is orthogonal [1], for which \(\mathbf {H}_{k}\mathbf {H}_{k}^{H}\) is a diagonal matrix with \(||\mathbf {H}_{k,m}||^{2}_{F}\) as its (m,m)-th element, where \(||\mathbf {H}_{k,m}||^{2}_{F}\) is the squared Frobenius norm of the mth row of H _{ k }. In this case, \(\gamma _{m}\mathbf {(H}_{k}) = ||\mathbf {H}_{k,m}||^{2}_{F}\). The MIMO channel capacity is calculated at each OFDM subcarrier using the above equation, while the MIMO-OFDM channel capacity is calculated as an average of the MIMO channel capacity over all OFDM subcarriers at a fixed signal SNR = 20 dB. For our measurement analysis, we chose signal SNR = 20 dB to have an even estimation of the MIMO channel capacity along the path and to emphasize effects of the MIMO channel structure on the capacity. The fixed value of signal SNR = 20 dB was chosen as an example, which is typically used for MIMO channel capacity analysis for the system using 16-quadrature amplitude modulation (QAM) or 64-QAM (e.g., [27, 37])
We note that the MIMO capacity value monotonically increases or decreases with the higher or lower signal SNR and it does not affect the conclusions of comparative analysis of higher or lower MIMO channel capacity as a function of subcarrier spacing or environment as have been performed in this paper. The MIMO channel capacity referred to in this paper corresponds to an ideal capacity, where it is assumed that a perfect CSI is available at the receiver. The actual capacity is typically reduced from this capacity due to the inaccuracy of the CSI at the receiver, especially in a mobile environment. In addition, the effects of inter-carrier interference (ICI) and inter-symbol interference (ISI) are ignored in this paper. We consider that the effects of ICI are small given the subcarrier spacing used and the maximum Doppler shift assumed. The ISI is expected to be removed by the use of cyclic prefix.
For a SISO fading channel with perfect knowledge of the channel at the receiver, the capacity of a SISO link is given as C= log2(1+ρ). Therefore, the capacity of an equivalent SISO link is equal to approximately 6.66 bps/Hz at ρ.
In this paper, there are two types of SNR being discussed. The first is nominated as ρ in (2) and is referred as signal SNR. The second SNR is referred to as channel estimation (CE) SNR which indicates the quality of the channel estimation.
1.2.3 1.2.3 Maximum LOS MIMO capacity criteria
Conceptually, a larger TX-RX separation distance requires larger antenna element spacings, and lower frequencies require larger antenna spacings. For fixed f and D, the antenna arrays could be easily designed so that subchannel orthogonality can be achieved which will result to the attainment of maximum capacity in LOS scenarios. For V2I communications where the RX vehicle is moving, D and θ change with time. Hence, the receiver antenna array is not fixed at a specific position, but its location varies with the motion of the vehicle. The optimal LOS MIMO V2I operation depends on achieving the optimal angle θ _{opt} and separation distance between TX and RX D _{opt}.
1.2.4 1.2.4 Least square channel estimation
It is important to note that this simple LS estimate \(\hat {H}_{\text {\tiny {LS}}}\) does not exploit the correlation of channels across frequency carriers and across OFDM symbols. We utilized the LS estimation approach to get the initial MIMO channel estimates at the pilot subcarriers, which was then further improved using a 2D DCT filtering technique in both time and frequency [30–32]. The effectiveness of this technique is shown in the Section 1.4.
1.3 MIMO-OFDM V2I measurements
1.3.1 1.3.1 Measurement equipment
OFDM packet parameters
Parameters | LSS | MSS | SSS |
---|---|---|---|
Baseband sample rate (Msps) | 32.768 | 32.768 | 32.768 |
Measurement channel bandwidth (MHz) | 8.2 | 9.0 | 6.7 |
Center frequency (MHz) | 920 | 920 | 920 |
Modulation | QPSK | QPSK | QPSK |
No. of OFDM symbols | 6003 | 593 | 156 |
No. of occupied subcarriers | 52 | 600 | 1705 |
Subcarrier spacing (kHz) | 156.25 | 15 | 3.348 |
FFT length | 208 | 2184 | 8384 |
OFDM symbol period (μs) | 6.4 | 66.67 | 299 |
Guard interval (μs) | 1.6 | 16.67 | 75 |
A custom-built user interface software running on Windows operating system was used. This software captures ADC samples and records them on a PC’s hard disk via USB connection. The transmitter sends OFDM signals to sound the channel, which are eventually recorded at the receiver unit. By post-processing, we then obtained the complex channel transfer function or frequency response as explained in Section 1.2.4. The transmitted OFDM signal waveforms (DAC samples) were generated off-line in the PC, using MATLAB. The in-house built platform is equipped with a field-programmable gate array (FPGA) that streams the off-line generated waveforms to the AD9361’s DAC. The baseband-to-RF module amplifies, filters, and up-converts the signals to RF (920 MHz), and the antenna module processes the signals from the RF stage and sends them to the receiver through the MIMO channel. At the receiver side, the signals reach the antennas and pass to the low-noise amplifiers and down converters. The baseband waveform is sampled by the ADC, and the sampled baseband waveform data is transferred to the PC via USB and recorded on the hard disk. Each transmitted OFDM-based wireless packet has an approximate duration of 50 ms. The transmitted OFDM packets were sampled at 32.768 M samples per second. The measurement channel bandwidths are 8.2, 9.0, and 6.7 MHz for LSS, MSS, and SSS, respectively. The transmitted data are modulated by quadrature phase shift keying (QPSK). The OFDM-based wireless packets entirely consisted of the pilot symbols for the purpose of the 2×2 MIMO-OFDM channel measurement.
A 2×2 MIMO antenna system was implemented using two commercially available omnidirectional vertically polarized L-COM HGV-906 antennas and two off-the-shelf omnidirectional antennas as both TX and RX antenna array elements, respectively, for all measurements described in this paper. A 35-dBm high-power amplifier ZHL-1000-3W from Mini-Circuits was added at the transmitter to provide an output power of 23 dBm. With the use of a 6-dBi antenna, the maximum effective isotropic radiated power (EIRP) is 29 dBm. The spacing of the antenna elements is set to ≈ 6 λ at TX and ≈ 3 λ at RX where λ=32.5 cm, i.e., the spacing between the TX antennas TX1-TX2=2 m and between the RX antennas is RX1-RX2=1 m. The two transmitter antennas (TX) were mounted at a fixed position each at the same height of H _{TX}=3.6 m and the two receiver antennas (RX) mounted at the rooftop of a vehicle at a height of H _{RX}=1.8 m, as shown in Fig. 3. The measurement platform uses built-in Global Positioning System (GPS) receivers to ensure accurate synchronization of the TX and RX. In addition, the receiver system was equipped with an external EVK-6T-0 U-blox 6 GPS to provide measurement time stamp as well as the location and speed data for the RX vehicle.
Measurement parameters
Parameters | Values |
---|---|
MIMO system | 2×2 |
Center frequency | 920 MHz |
Packet duration | 50 ms |
Sampling frequency | 32.768 Msps |
Transmit power | 23 dBm |
TX antenna gain | 6 dBi |
RX antenna gain | 2 dBi |
TX antenna spacing | 2 m |
RX antenna spacing | 1 m |
TX antenna height, h _{TX} | 3.6 m |
RX antenna height, h _{RX} | 1.8 m |
Modulation | QPSK |
1.3.2 1.3.2 Measurement environments
1.4 Measurement results and analysis
The post-processing of measurement data was performed using MATLAB. The measured MIMO channels are estimated from the received signals. In particular, the channel coefficients were estimated by LS channel estimation by sending known channel training symbols alternately in time using the two transmitters (e.g., the first transmitter sent channel training symbols at odd OFDM symbol time index and the second transmitter sent channel training symbols at even OFDM symbol time index). The transmitted wireless packets have a duration of 50 ms and consist of all known channel training symbols.
The measurement system was equipped with GPS receivers at both the TX and the RX ends for accurate time reference (less than 100 ns ambiguity), which facilitated frame detection and time offset estimation. The GPS receivers also provided accurate frequency reference which removed the necessity of frequency offset estimation and correction. Residual frequency offset which was less than 100 Hz was included in the estimated channel. No additional ICI cancelation technique was applied, apart from the fact that the TX and the RX frequency references were synchronized to the GPS reference, which we expect to reduce the ICI problem significantly. We note that the Doppler shift expected in the measurement at 920 MHz with up to 55 km/h is approximately 47 Hz. Given the minimum subcarrier spacing of approximately 4 kHz (corresponding to SSS), we consider that the effects are negligible.
We used a high sampling rate of 32.768 Msps for the measurement to reduce the effect of noise. The use of oversampling to improve the SNR in OFDM transmission has been proposed in the literature [40, 41]. Random receiver noise is reduced by oversampling based on the assumption that the signal is coherent and that the noise is random. We extended our analysis to remove the effects of channel estimation errors by applying 2D DCT filtering in both time and frequency [30–32]. The 2D DCT has significantly reduced the effect of noise on our measurement results. In addition, we have included a simulation analysis to show that the effects of channel estimation error on MIMO channel capacity are small for high SNR up to 20 dB.
1.4.1 1.4.1 Effects of channel estimation error on MIMO channel capacity
1.4.2 1.4.2 Measured MIMO-OFDM channel capacity variation as a function of the measurement time
1.4.3 1.4.3 Cumulative distribution function of MIMO capacity
Comparison of the CDF of MIMO capacity for all the three subcarrier spacing scenarios, C (in bps/Hz)
Standards | C _{5 % } | C _{10 % } | C _{50 % } | C _{90 % } | C _{95 % } | C _{DR(90 %)} |
---|---|---|---|---|---|---|
LSS | 9.2 | 9.7 | 11.4 | 12.2 | 12.4 | 3.2 |
MSS | 9.2 | 9.7 | 11.4 | 12.5 | 12.8 | 3.6 |
SSS | 9.2 | 9.7 | 11.4 | 13.0 | 13.5 | 4.3 |
1.4.4 1.4.4 MIMO capacity as a function of time and frequency
1.4.5 1.4.5 MIMO channel capacity against separation distance
In general, Fig. 22 clearly shows a correlation between the MIMO channel capacity variation and the TX-RX separation distance, i.e., MIMO channel capacity has a larger variation in near-range LOS conditions and a smaller capacity variation in far-range NLOS conditions. The possible cause of the larger capacity variations for near-range LOS propagation is greater angular spacing between the MIMO subchannels. There is greater angular spacing between the MIMO subchannels for LOS links and less angular spacing between the MIMO subchannels for far-range NLOS links, as illustrated in Fig. 2. Figure 2 shows that as the degree of angular spacing between the MIMO subchannels becomes smaller, the correlation between the received signals becomes larger, as the differences in the length of the MIMO subchannel paths decrease. This greater angular spacing causes the different MIMO propagation paths d _{11} and d _{12}, as well as d _{21} and d _{22}, to be orthogonal, therefore increasing the expected channel capacity. The figure explains our rationale to relate the degree of angular spacing between the MIMO subchannels and the correlation between the received signals. Most LOS conditions in our measurements met the criteria for maximum capacity as stated in Eq. (6), which indicates that the propagation paths should meet the conditions d _{11} and d _{12} and d _{21} and d _{22} that occur when both the TX and RX arrays are parallel. These conditions are met in most of the measured LOS scenarios discussed in this paper. The results presented in this section are consistent with what is expected, as they show that the mean MIMO capacity varies according to the degree of angular spacing between the MIMO subchannels and the correlation between the received signals.
It is observed that the maximum capacities for all subcarrier spacing scenarios were achieved by the LOS case despite the assumed highly correlated channel. This is a surprising result, because signal SNR is not a function of location in our analysis since we considered a fixed signal SNR = 20 dB in order to separate the effects of signal SNR variations. In this measurement, TX-RX antenna arrays are separated far apart on each side of the link (TX =6λ and RX =3λ) which is considered to contribute to the de-correlation of the MIMO subchannels in near-range LOS scenarios and to result in a full-rank MIMO channel and lead to the achievement of maximum MIMO capacity in LOS scenarios. The results are consistent with the phenomenon known as LOS MIMO in the literature [14, 24, 25, 42].
Based on the maximum LOS MIMO capacity criteria presented in the previous sections, we investigated a 2×2 MIMO V2I system operating in LOS and NLOS conditions. The TX-RX separation distance D varies between 3 and 250 m as the RX vehicle drives away from TX at a carrier frequency f=920 MHz and fixed antenna array spacing s _{1}=2 m at TX and s _{2}=1 m at RX; θ and D _{opt} varies as the RX vehicle drives around the vicinity. At a TX height of 3.6 m and a RX height of 1.8 m, the calculated optimal LOS TX-RX separation distance D _{opt} that will satisfy (9) for r=0 and θ=40° is D _{opt}=7.8 m.
In our MIMO-OFDM MSS case, the measured C _{max}≈ 13.0 bps/Hz which is the same as the theoretical maximum LOS-MIMO capacity C _{max}=13.0 bps/Hz that was achieved at a LOS path corresponding to the theoretical D _{opt}=13.1 m, which satisfies (9) for fixed f=920 MHz, s _{1}=2 m, s _{2}=1 m, θ=0, and r=0. While for the case of LSS and SSS, a higher value than the theoretical maximum capacity of C _{max}=13.37 bps/Hz and C _{max}=13.57 bps/Hz, respectively, was achieved at LOS distances of D=76.7 m and D=62 m, respectively. This shows that our antenna spacing satisfied the criterion for achieving the maximum MIMO capacity in LOS situations by proper positioning or spacing of the antenna elements in such a way to achieve orthogonality between spatially multiplexed signals of MIMO systems, resulting in a high-rank channel response matrix. In conclusion, all the maximum capacity values were achieved at near-range LOS and the measured capacity shows very close agreement with the theoretical values of optimal LOS MIMO capacity with very small variations which we attribute to minor inaccuracies in the positioning and orientation of the antenna elements. This validates the LOS maximum capacity criterion for achieving orthogonality between spatially multiplexed signals of MIMO systems in LOS channels by employing specifically designed antenna arrays.
2 Conclusions
A novel 2×2 MIMO testbed has been designed and implemented for three different subcarrier spacings. The MIMO testbed is based on a software-defined radio platform where flexible signal processing can be implemented. The testbed has been used to carry out measurements at a frequency of 920 MHz. We have investigated the channel capacity of a MIMO-OFDM V2I channel under both LOS and NLOS propagation scenarios. We examined a method to achieve orthogonality between spatially multiplexed signals in MIMO V2I communication systems operating in a LOS channel. Our analytical results show that the maximum capacity can be achieved under LOS scenario with proper antenna element spacing despite the effects of higher correlation and reduced rank of the channel response matrix which can be counterbalanced by deliberate separation of antenna elements that preserves orthogonality; this results in a full-rank MIMO channel matrix, and thus, high MIMO capacity is achieved in LOS cases. As anticipated, when the vehicle is driving closer to the TX roadside infrastructure unit, the greatest capacity values are observed in LOS conditions. We observed that even with some deviation from optimal design, the LOS MIMO case outperforms the theoretical i.i.d. Rayleigh performance (11.5 bps/Hz) in terms of Shannon capacity. We investigated the MIMO channel capacity for different positions of the receiver vehicle. Interestingly, since the maximum capacity of LOS-MIMO systems does not depend on the existence of scatterers, there is no particular limit on the linear increase of the system capacity. The presented results demonstrate that the MIMO channel benefits from the movement of the receiver from NLOS to LOS conditions, as the detrimental effect of increased correlation between the received signal is outweighed by the advantageous effect of high angular spacing of the sub-paths. These results show the strong dependence of the MIMO capacity on antenna array spacing and the correlation of the channel response matrix. This measurement demonstrates the significance of using optimally interspaced MIMO antenna array elements at each of the radio links to increase the capacity of MIMO systems over SISO systems.
Declarations
Acknowledgements
The authors would like to thank the QUT High Performance Computing (HPC) and the Research Support services for their technical assistance during the data analysis. The authors thank Alex Grancea and Joseph Pathikulangara from CSIRO ICT Centre NSW, Australia, for the development of the SDR platform used for the measurements.
Authors’ Affiliations
References
- G Foschini, M Gans, On limits of wireless communications in a fading environment when using multiple antennas. Wirel. Pers. Commun. 10(6), 311–335 (1998).View ArticleGoogle Scholar
- H Cheng, H Shan, W Zhuang, Infotainment and road safety service support in vehicular networking. Elsevier J. Mech. Syst. Signal Process. 2010, 1–19 (2010).View ArticleGoogle Scholar
- W K Ziri-Castro, N Scanlon, Evans, Prediction of variation in MIMO channel capacity for the populated indoor environment using a radar cross-section-based pedestrian model. IEEE Trans. Wirel. Commun. 4(3), 1186–1194 (2005).View ArticleGoogle Scholar
- H Suzuki, T Van, I Collings, Characteristics of MIMO-OFDM channels in indoor environments. EURASIP J. Wirel. Commun. Netw. 2007(19728), 1–9 (2007).View ArticleGoogle Scholar
- A Goldsmith, S Jafar, N Jindal, S Vishwanath, Capacity limits of MIMO channels. IEEE J. Sel. Areas Commun. 21(5), 684–702 (2003).View ArticleGoogle Scholar
- A Molisch, M Steinbauer, M Toeltsch, E Bonek, R Thoma, Capacity of MIMO systems based on measured wireless channels. IEEE J. Sel. Areas Commun. 20(3), 561–569 (2002).View ArticleGoogle Scholar
- GXu Y Yang, H Ling, An experimental investigation of wideband MIMO channel characteristics based on outdoor non-LOS measurements at 1.8 GHz. IEEE Trans. Antennas Propag. 54(11), 3274–3284 (2006).View ArticleGoogle Scholar
- J Karedal, F Tufvesson, N Czink, A Paier, C Dumard, T Zemen, C Mecklenbrauker, A Molisch, A geometry-based stochastic MIMO model for vehicle-to-vehicle communications. IEEE Trans. Wirel. Commun. 8(7), 3646–3657 (2009).View ArticleGoogle Scholar
- J Ling, D Chizhik, D Samardzija, R Valenzuela, Peer-to-peer MIMO radio channel measurements in a rural area. 6(9), 3229–3237 (2007).Google Scholar
- J Karedal, F Tufvesson, N Czink, A Paier, C Dumard, T Zemen, C Mecklenbrauker, A Molisch, in Communications, 2009. ICC ‘09. IEEE International Conference on. Measurement-based modeling of vehicle-to-vehicle MIMO channels (Dresden, 14–18 June 2009), pp. 1–6.Google Scholar
- O Renaudin, V Kolmonen, P Vainikainen, C Oestges, Non-stationary narrowband MIMO inter-vehicle channel characterization in the 5 GHz band. IEEE Trans. Veh. Technol. 59(4), 2007–2015 (2010).View ArticleGoogle Scholar
- A Alonso, A Paier, T Zemen, N Czink, F Tufvesson, in Communications Workshops (ICC), 2010 IEEE International Conference on. Capacity evaluation of measured vehicle-to-vehicle radio channels at 5.2 GHz (Cape Town, 23–27 May 2010), pp. 1–5.Google Scholar
- H Mousavi, B Khalighinejad, B Khalaj, in Wireless Communications and Mobile Computing Conference (IWCMC), 2013 9th International. Capacity maximization in MIMO vehicular communication using a novel antenna selection algorithm (Sardinia, 1–5 July 2013), pp. 1246–1251.Google Scholar
- M Matthaiou, D Laurenson, C-X Wang, in Wireless Communications and Networking Conference, 2008. WCNC 2008 IEEE. Capacity study of vehicle-to-roadside MIMO channels with a line-of-sight component (Las Vegas, NV, 3 April 2008 31 March), pp. 775–779.Google Scholar
- A Paier, J Karedal, N Czink, H Hofstetter, C Dumard, T Zemen, F Tufvesson, C Mecklenbrauker, A Molisch, in Personal, Indoor and Mobile Radio Communications, 2007. PIMRC 2007. IEEE 18th International Symposium on. First results from car-to-car and car-to-infrastructure radio channel measurements at 5.2 GHz (Athens, 3–7 September 2007), pp. 1–5.Google Scholar
- L Cheng, B Henty, D Stancil, F Bai, P Mudalige, Mobile vehicle-to-vehicle narrow-band channel measurement and characterization of the 5.9 GHz dedicated short range communication (DSRC) frequency band. IEEE J. Sel. Areas Commun. 25(8), 1501–1516 (2007).View ArticleGoogle Scholar
- G Grant, Rayleigh fading multi-antenna channels. J. Appl. Signal Process. 2002(3), 316–329 (2002).MATHMathSciNetView ArticleGoogle Scholar
- H Shin, JH Lee, Capacity of multiple-antenna fading channels: spatial fading correlation, double scattering, and keyhole. IEEE Trans. Inf. Theory. 49(10), 2636–2647 (2003).MATHMathSciNetView ArticleGoogle Scholar
- M Chiani, M Win, A Zanella, On the capacity of spatially correlated MIMO Rayleigh-fading channels. IEEE Trans. Inf. Theory. 49(10), 2363–2371 (2003).MATHMathSciNetView ArticleGoogle Scholar
- S Jin, X Gao, X You, On the ergodic capacity of rank-1 Ricean-fading MIMO channels. IEEE Trans. Inf. Theory. 53(2), 502–517 (2007).MATHMathSciNetView ArticleGoogle Scholar
- M Kang, M-S Alouini, Capacity of MIMO Rician channels. IEEE Trans. Wirel. Commun. 5(1), 112–122 (2006).View ArticleGoogle Scholar
- P Driessen, G Foschini, On the capacity formula for multiple input-multiple output wireless channels: a geometric interpretation. IEEE Trans. Commun. 47(2), 173–176 (1999).View ArticleGoogle Scholar
- D Gesbert, DGORE H Bolcskei, A Paulraj, Outdoor MIMO wireless channels: models and performance prediction. IEEE Trans. Commun. 50(12), 1926–1934 (2002).View ArticleGoogle Scholar
- M Matthaiou, P de Kerret, G Karagiannidis, J Nossek, Mutual information statistics and beamforming performance analysis of optimized LOS MIMO systems. IEEE Trans. Commun. 58(11), 3316–3329 (2010).View ArticleGoogle Scholar
- I Sarris, A Nix, Design and performance assessment of maximum capacity MIMO architectures in line-of-sight. IEEE Proc. Commun. 153(4), 482–488 (2006).View ArticleGoogle Scholar
- F Bohagen, P Orten, G Oien, in Wireless Communications and Networking Conference, 2005 IEEE, 1. Construction and capacity analysis of high-rank line-of-sight MIMO channels (New Orleans, Louisiana, USA, 13–17 March 2005), pp. 432–437.Google Scholar
- I Sarris, A Nix, Design and performance assessment of high-capacity MIMO architectures in the presence of a line-of-sight component. IEEE Trans. Veh. Technol. 56(4), 2194–2202 (2007).View ArticleGoogle Scholar
- F Bohagen, P Orten, G Oien, Design of optimal high-rank line-of-sight MIMO channels. IEEE Trans. Wirel. Commun. 6(4), 1420–1425 (2007).View ArticleGoogle Scholar
- M Ozdemir, H Arslan, Channel estimation for wireless OFDM systems. IEEE Commun. Surv. Tutorials. 9(2), 18–48 (2007).View ArticleGoogle Scholar
- P Tan, N Beaulieu, A comparison of DCT-based FDM and DFT-based OFDM in frequency offset and fading channels. IEEE Trans. Commun. 54(11), 2113–2125 (2006).View ArticleGoogle Scholar
- H N Al-Dhahir, S Minn, Satish, Optimum DCT-based multicarrier transceivers for frequency-selective channels. IEEE Trans. Commun. 54(5), 911–921 (2006).View ArticleGoogle Scholar
- L Han, B Wu, in Wireless Communications and Signal Processing (WCSP), 2010 International Conference on. DCT-based channel estimation for wireless OFDM systems with the hexagonal pilot pattern (Suzhou, China, 21–23 October 2010), pp. 1–5.Google Scholar
- F Farrokhi, A Lozano, G Foschini, in Personal, Indoor and Mobile Radio Communications, 2000. PIMRC 2000. The 11th IEEE International Symposium on, 1. Spectral efficiency of wireless systems with multiple transmit and receive antennas (London, 373).Google Scholar
- F Farrokhi, A Lozano, G Foschini, R Valenzuela, Spectral efficiency of FDMA/TDMA wireless systems with transmit and receive antenna arrays. IEEE Trans. Wirel. Commun. 1(4), 591–599 (2002).View ArticleGoogle Scholar
- C Shannon, A mathematical theory of communication. Bell Syst. Technol. 27, 379–423 (1948).MATHMathSciNetView ArticleGoogle Scholar
- J Winters, On the capacity of radio communication systems with diversity in a Rayleigh fading environment. IEEE J. Sel. Areas Commun. 5(5), 871–878 (1987).View ArticleGoogle Scholar
- A Toding, M Khandaker, Y Rong, Joint source and relay design for MIMO multi-relays systems using projected gradient approach. EURASIP J. Wirel. Commun. Netw. 2014(155), 1–22 (2014).Google Scholar
- J Kermoal, L Schumacher, K Pedersen, P Mogensen, F Frederiksen, A stochastic MIMO radio channel model with experimental validation. IEEE J. Selected Areas Commun. 20(6), 1211–1226 (2002).View ArticleGoogle Scholar
- P Kyritsi, D Chizhik, Capacity of multiple antenna systems in free space and above perfect ground. IEEE Commun. Lett. 6(8), 325–327 (2002).View ArticleGoogle Scholar
- J Wu, Y Zheng, Oversampled orthogonal frequency division multiplexing in doubly selective fading channels. IEEE Trans. Commun. 59(3), 815–822 (2011).View ArticleGoogle Scholar
- B-S Y-S Lee, OFDM Seo, receivers using oversampling with rational sampling ratios. IEEE Trans. Consum. Electron. 55(4), 1765–1770 (2009).View ArticleGoogle Scholar
- D McNamara, M Beach, P Fletcher, P Karlsson, Capacity variation of indoor multiple-input multiple-output channels. Electron. Lett. 36(24), 2037–2038 (2000).View ArticleGoogle Scholar
Copyright
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.