Chapter 3: Measurement System and Environment
3.2 Measurement Setup
For the measurements, we used the wideband vector channel sounder RUSK ATM with a measurement bandwidth of 120 MHz at a center frequency of 2.44 GHz. At the receive side a λ/2 spaced 8-element uniform linear patch array (ULA) with two
Chapter 3 Measurement System and Environment
additional dummy elements was used. Each single patch antenna had a 3dB beamwidth of 120 degree and was consecutively multiplexed a single receiver chain. At the transmit side, a omni-directional antenna was moved along the measured path, we can form 8 virtual Tx antenna array without mutual coupling.
The receiving antenna was mounted on a rooftop at 2.44GHz with the transmission power of 1W. The transmitter antenna was carried in a trolley and was 1.8 meter above the road. In order to get multipath components, we sampled data by fixed-point and moving measurements along selected routes with walking speed
Figure 3-1 System diagram of the RUSK Channel Sounder
Chapter 3 Measurement System and Environment
Figure 3-2 pictures of the RUSK Channel Sounder (a) transmitter with an omnidirectional antenna;
and (b) the receiver with 8-element linear antenna
3.3 Measurement Environment Description
The measurement was performed in 9th floor (site A), 2nd floor (site B), 3st floor (site C), 4st floor (site D) and 5st floor (site E) of the 4th Engineering Building at the National Chiao-Tung University, Hsinchu, Taiwan and the layout is shown in Figs. 3-3 to 3-6.
At site A, path 1 and path2 is measured in the corridor, thus LOS always exists in path 1, instead LOS, the NLOS exists in path 2.At site B path1 is the LOS environment and the distance is longer than the site A, we will observe how the long distance affect the capacity.
Site D and site E is measured also in the hallway, but at the site E, the wall is constructed of the glasses and at site D is concrete wall. We will explore the different material how to affect the capacity.
Chapter 3 Measurement System and Environment
Figure 3-3 Floor layout of site A located on the second floor of Engineering Building No4
Chapter 3 Measurement System and Environment
Figure 3-4 Floor layout of site B located on the second floor of Engineering Building No4
Chapter 3 Measurement System and Environment
Figure 3-5 Floor layout of site C located on the second floor of Engineering Building No4
Chapter 3 Measurement System and Environment
Figure 3-6 Floor layout of site D located on the second floor of Engineering Building No4
Chapter 3 Measurement System and Environment
Table 2 Measurement situation
hallway Frequency Power Bandwidth Sample point Distance(m) Path1 2.44GHz 1 W 120 MHz 21900 24
Site A
Path2 2.44GHz 1 W 120 MHz 43000 53 Site B Path1 2.44GHz 1 W 120 MHz 26550 42
Site C Path1 2.44GHz 1 W 120 MHz 22100 48
Site D Path1 2.44GHz 1 W 120 MHz 23500 48
Chapter 4 Effects of Array Element Spacing and Multipath Propagation on MIMO Capacity in Indoors
Chapter 4
Effects of Array Element Spacing and Multipath Propagation on MIMO Capacity in Indoors
In this chapter, the effects of transmitting/receiving array spacing, propagation conditions (LOS and NLOS), and local scatterer distribution and signal bandwidth on MIMO capacity are investigated through the measurement. The complex correlation coefficients of the MIMO channel frequency response will be introduced and evaluated.
4.1 MIMO capacity evaluation
From Eq. (2-1), the 8x8 MIMO capacity is given by
2 8
( ) log (det( ( ) ( ) )) 8
C f I γ H f H f H
= +
The capacity is calculated withγ =30 dB and the measured 8x8 MIMO channel matrix, H, which is realized through the measurement by the RUSK channel sounder systems.
During the measurement, a virtual 8-element linear array at transmitting site is formed by grouping 8 associated Tx positions with chosen neighboring-position spacing.
From equation (2-3), the spatial correlation coefficient at the transmitter between elements is calculated.
Chapter 4 Effects of Array Element Spacing and Multipath Propagation on MIMO Capacity in Indoors
4.2. Propagation distance and array-element spacing effects
To investigate propagation and the element spacing effects on the MIMO capacity and Tx-Rx element correlation, the measurement routes have been selected to consider LOS and NLOS scenarios at Site A.
A. Propagation range effect
Figures 4-1 and 4-2 illustrate the MIMO capacity versus Tx-Rx distance for LOS and NLOS scenarios, respectively. In each figure, the array element spacing, Δ, is equal to 0.1 λ, 0.2λ,…., or 1.0λ. The figures show that the capacity is decreased as the Tx-Rx distance increases for both LOS and NLOS cases. It is because that when the Tx-Rx distance decreases, the rms angle spread of AOA increases, which reduces the correlation between the Tx and Rx elements, i.e., enhances the capacity. In the next section, we will analyze the relation of angle spread and MIMO capacity. Figures 4-3 and 4-4 show the correlation coefficient versus propagation distance for LOS and NLOS scenarios, respectively. The spatial correlation coefficient increases as the Tx-Rx distance increases.
B. Element spacing effect
From figures 4-1 to 4-4, it is also found that larger element spacing leads to lower correlation coefficient, i.e., higher MIMO capacity.
To further investigate the effect of element spacing, figures 4-5 and 4-6 illustrate capacity versus element spacing for LOS and NLOS scenarios, respectively. It is found that the capacity increases as the element spacing increases and it saturates when the spacing is larger than one wavelength. This reveals that the correlation distance between the elements in indoor environments is about one wavelength.
Chapter 4 Effects of Array Element Spacing and Multipath Propagation on MIMO Capacity in Indoors
Figure 4-1 Capacity versus Tx-Rx distance (d) for different Tx element spacing (LOS)
5 10 15 20 25 30 35 40 45 50
Figure 4-1 Capacity versus Tx-Rx distance (d) for different Tx element spacing (NLOS)
Chapter 4 Effects of Array Element Spacing and Multipath Propagation on MIMO Capacity in Indoors
Figure 4-3 Correlation coefficient versus d for different Tx element spacing (LOS)
Figure 4-3 Correlation coefficient versus d for different Tx element spacing (NLOS)
Chapter 4 Effects of Array Element Spacing and Multipath Propagation on MIMO Capacity in Indoors
Figure 4-5 Capacity versus transmitter antenna element spacing for LOS
0 .5 1 1 .5 2 2 . 5 3 3 .5 4 4 .5 5
A ntenna elem ent spac ing
Capacity
Tx-Rx distance= 1m Tx-Rx distance= 15m
Figure 4-6 Capacity versus transmitter antenna element spacing for NLOS
Chapter 4 Effects of Array Element Spacing and Multipath Propagation on MIMO Capacity in Indoors
C. Effect of unequal element spacing
In the MIMO system, channel capacity may be improved by adaptively changing the element spacing. Changing the element spacing is a way to provide spatial diversity to a MIMO link without increasing the number of antenna array elements [4]. In this thesis, we consider adapting the location of elements at both ends of the link. Here, we adjust the element spacing of the virtual antenna arrays at the transmitter to investigate how the capacity varies with small changes in element locations.
To investigate how the channel capacity changes with element spacing, and how much benefit we can get by the adaptive array. From measured data of site A, we synthesized virtual 141-element uniform linear arrays, whose neighboring element spacing is 0.1λ. At the receiver the number of antenna arrays is fixed to 8 and the antenna array spacing is 0.4λ (Following the RUSK sounder specification). The channel matrices of unequally spaced array elements can be determined as subsets of the measured (141,8) MIMO channel. We randomly choose 8 element at transmitter to form an 8×8 MIMO without mutual coupling. Figures 4-7 and 4-8 demonstrate MIMO capacities of unequal- and equal- spacing arrays for LOS and NLOS cases, respectively.
The element spacing of the equally spaced array is equal to 2λ. The figures show that the unequal spacing arrays always lead to an optimum capacity, which is better than that of the equal spacing arrays.
The reference [4] mentioned that if both transmitter antenna array spacing and receiver antenna array spacing is unequal; the capacity will reach the maximum. Now we also change the transmitter array spacing and receiver array spacing. Because the receiver antenna array is fixed to 8 elements, a 4×4 MIMO system is chosen to have an allowance to change the receiver array spacing. From figures 4-9 and 4-10, we can observe that
Chapter 4 Effects of Array Element Spacing and Multipath Propagation on MIMO Capacity in Indoors
when the spacing of transmitter array and receiver array is unequal, the capacity may reach maximum. The reference [4] explains that the correlation of unequal antenna spacing is lower than equal antenna spacing. So lower correlation coefficient result in high capacity. Figures 4-11 and 4-12 illustrate the capacity versus Tx standard deviation for LOS and NLOS, respectively. The standard deviation of element spacing is expressed
as
( )
2 121
1 N
i i
a a σ ∆ N
=
= ⎡⎢⎣
∑
− ⎦⎤⎥ , where N is the number of antenna array; ai is element array spacing.The measured results show that the MIMO capacity is optimized When the standard deviation of Tx element spacing is ranging from 1λ to 1.5λ.Chapter 4 Effects of Array Element Spacing and Multipath Propagation on MIMO Capacity in Indoors
Figure 4-7 Capacity of equal antenna array and unequal antenna array as a function of d (Site A, path1)
5 10 15 20 25 30 35 40 45 50
20 22 24 26 28 30 32
distance(m)
capacity
equally spaced array unequally spaced array with max. capacity
Figure 4-8 Capacity of equal antenna array and unequal antenna array as a function of d (Site A, path2)
Chapter 4 Effects of Array Element Spacing and Multipath Propagation on MIMO Capacity in Indoors
Figure 4-9 4×4 MIMO system capacity for equal array and unequal array (Site A,path1)
2 4 6 8 10 12 14 16 18 20 max. capacity by moving Tx and Rx
Figure 4-10 4×4 MIMO system capacity for equal array and unequal array (Site A,path2)
Chapter 4 Effects of Array Element Spacing and Multipath Propagation on MIMO Capacity in Indoors
σ∆ σ∆
(a) (b)
(b)
σ∆ σ∆
(c) (d)
Figure 4-11 σ∆ for site A (LOS). In each figure, the maximum value of the capacity is located and its corresponding σ∆ is shown as a mark *.
(a) Tx-Rx distance: 4m (b) Tx-Rx distance:10m (c) Tx-Rx distance:15m (d)
Tx-Rx distance: 20m
Chapter 4 Effects of Array Element Spacing and Multipath Propagation on MIMO Capacity in Indoors
(b)
Figure 4-11
(a)
σ∆
σ∆
σ∆ σ∆
(c) (d)
σ∆ for site A (LOS). In each figure, the maximum value of the capacity is located and its corresponding σ∆ is shown as a mark *.
(a)Tx-Rx distance: 16m (b) Tx-Rx distance:24m (c) Tx-Rx distance:37m
d)
( Tx-Rx distance: 41m
Chapter 4 Eff ulti Capacity in Indoors
4.3 Angle
To investigate ent effect on capacity and spatial correlations, we will introduce ang
or the large values of the angle spread will
y decreases along the hallway, but at the distance from 10m to 15m, we
e at figure 3-5 and
ects of Array Element Spacing and M path Propagation on MIMO
spread effect
the propagation environm
le spread to observe the relation of capacity and propagation distance. The reference [12] mentions that f
result in low correlation coefficient, low correlation coefficient obtains high capacity. We will observe the relation of capacity and angle spread during the measurement of site B, site C and site D.
At site B, the measurement carries out in the 2nd floor of the 4th Engineering Building. The measured path and Rx position illustrates in figure 3-3. From figure 4-13, it shows that capacit
find that the capacity goes down suddenly. In order to explain this phenomenon, we can observe the measured map from figure 3-3, a lobby locates at the measured route from 10m to 15m and the lobby is an open area. We guess that in this open area, the angle spread decrease suddenly and the channel becomes correlated, so the capacity goes down suddenly. At figure 4-14, we plot the angle spread versus the propagation distance. We find that angle spread also decrease when the Tx-Rx distance increase. At 10m to 15m, angle spread also decrease suddenly. The measured result is as same as and reference [12], large angle spread may obtain more multipath. In multipath richness environment, the channel correlation will become lower and the capacity will be larger.
To prove the angle spread how affects the capacity. We carry out the same measurements at different floor. Site C and site D measured in 4th floor and 5th floor of the 4th Engineering Building, respectively. The measured route illustrat
figure 3-6. The 4th floor and 5th floor have the same structure, but in the 5th floor the wall is constructed of glass and the wall of 4th floor is made of concrete. We will observe the different material of wall how affects the capacity.
Chapter 4 Effects of Array Element Spacing and Multipath Propagation on MIMO Capacity in Indoors
From figure 4-15 and figure 4-17, we find that capacity also decrease when propagation distance increase. But under the same antenna array spacing, the capacity of site C is always larger than the site D. We guess that the wall of concrete will reflect more multipath than glass and angle spread of site C will also be larger than site D. Figures 4-16 and 4-18 show angle spread versus propagation distance for different floors. We find that in figure 4-16, angle spread is significant lager than figure 4-18. So under rich multipath and lager angle spread environments between each transmitter and receiver antenna pair, MIMO wireless communications systems achieve significant capacity gains over conventional single antenna systems.
Chapter 4 Effects of Array Element Spacing and Multipath Propagation on MIMO Capacity in Indoors
Figure 4-14 rms angle spread of AOA versus Tx-Rx distance at Site B Figure 4-13 Capacity versus d for different Tx element spacing at Site B
5 10 15 20 25 30 35
Chapter 4 Effects of Array Element Spacing and Multipath Propagation on MIMO Capacity in Indoors
Figure 4-15 Capacity versus d for different Tx element spacing at Site C
5 10 15 20 25 30 35 40 45
Figure 4-16 rms angle spread of AOA versus d at Site C
Chapter 4 Effects of Array Element Spacing and Multipath Propagation on MIMO Capacity in Indoors
Figure 4-17 Capacity versus d at Site D for different Tx element spacing
5 10 15 20 25 30 35 40 45
Figure 4-18 rms angle spread of AOA versus d at Site D for different Tx element spacing
Chapter 4 Effects of Array Element Spacing and Multipath Propagation on MIMO Capacity in Indoors
4.4 Bandwidth effect
In this section, we investigate the impact of the signal bandwidth on the MIMO capacity. We choose 120M and 20M signal bandwidths to see how it affects the capacity.
The experiment carried out in different floors (site C and site D). From figures 4-19 and 4-18, we can find that capacity of 120M-bandwidth is always greater than the 20M-bandwidth. Because the time resolution of 120M is 8.3ns and 20M is 49.8 ns, when bandwidth becomes large, time resolution decrease; hence array resolved more multipaths and channel becomes uncorrelated, uncorrelated channel results in high capacity.
Figures 4-21 and 4-23 illustrate multipath number versus distance; the result shows that multipath numbers of 120M-bandwidth is always greater than 20M-bandwidth; the
According to section 4-3, angle spread is also in proportion to capacity. On different bandwidth, we also observe th
, the angle spread almost measured results prove that high bandwidth will resolve more multipath numbers.
e variation of angle spread versus MIMO capacity. From figures 4-22 and 4-24, we find that on different bandwidth
similar. So the capacity bases on different bandwidth, the main effect are multipath numbers, in multipath richness environment, the capacity will be higher.
Chapter 4 Effects of Array Element Spacing and Multipath Propagation on MIMO Capacity in Indoors
Figure 4-19 Capacity versus d at site C for signal bandwidths of 20MHz and 120MHz
5 10 15 20 25 30 35
Figure 4-20 Capacity versus d at site D for signal bandwidths of 20MHz and 120MHz
Chapter 4 Effects of Array Element Spacing and Multipath Propagation on MIMO Capacity in Indoors
ber of multipath components at for signal bandwidths of 20MHz and
5 10 15 20 25 30 35
Figure 4-22 rms angle spread of AOA versus d for signal bandwidths of 20MHz and 120MHz (site C)
Chapter 4 Effects of Array Element Spacing and Multipath Propagation on MIMO Capacity in Indoors
Figure 4-23 Number of multipath signal bandwidths of 20MHz and 120MHz (site D)
Figure 4-24 rms angle spread of A ignal bandwidths of 20MHz and 120MHz (site D)
Chapter 4 Effects of Array Element Spacing and Multipath Propagation on MIMO Capacity in Indoors
4.5 Local scatterer effect
In this section, we analyze the local scatterer how affects the MIMO capacity. The figure 4-25 descried the measured environments. We measured at each receiving 9 points for each room, the broadside direction of receiving linear array points at three individual directions with 120° interval. Therefore, 3 MIMO capacity values are shown for each point. Figure 4-25 illustrates that direction of receiving array is not significant effect on capacity. At figure 4-26, we place two scatterers at each room and observe the variation of capacity, it shows that when the receiving array is directional to scatterer, the capacity is significant higher than others direction, it is because the scatterer reflect more
m .
But from figure 4-26, we also find that fect on
neighboring receiving array. At figure 4-27 we analyze how far the scatterer affects MIMO capacity. We measured at each receiving 12 points for each room, the broadside direction of receiving linear array point to scatterer. The interval distance of each receiver is 0.5m; figures 4-27 and 4-28 show MIMO capacity with scatterer and no scatterer condition, respectively. After comparing to measured results, we find that when the receiver within 1.5m from scatterer, MIMO capacity significant increases. In indoor environment, the effect distance of scatterer is about 1.5m.
ultipath , more multipath result in low correlation coefficient and obtain higher capacity scatterer only has significant ef
Chapter 4 Effects of Array Element Spacing and Multipath Propagation on MIMO Capacity in Indoors
Chapter 4 Effects of Array Element Spacing and Multipath Propagation on MIMO Capacity in Indoors
Chapter 4 Effects of Array Element Spacing and Multipath Propagation on MIMO Capacity in Indoors
Chapter 4 Effects of Array Element Spacing and Multipath Propagation on MIMO Capacity in Indoors
Chapter 5 Conclusion
Chapter 5
Conclusion
In this thesis the analysis of the impact of various conditions on 8x8 and 4x4 MIMO systems, capacity and correlations has been presented, including antenna spacing, angle spread, Tx-Rx distance, bandwidth and local scatter. The measurement using the RUSK channel sounder was carried out in the National Chiao Tung University campus. In this research, some phenomena are reveal and listed as following.
(1) Antenna array element spacing effect: antenna spacing may affect MIMO capacity and correlations significantly. The MIMO capacity increases as the element spacing increases and it saturates when the spacing is larger than one wavelength. This reveals that the correlation distance between the elements in indoor environments is about one wavelength.
It is also found that when the standard deviation of Tx element spacing is ranging from 1λ to 1.5λ. The capacity of the MIMO with unequal array spacing is always larger than equal spacing. (2) Propagation distance effect: MIMO capacity is also decreased as the Tx-Rx distance increases for both LOS and NLOS cases. It is because that when the Tx-Rx distance decreases, the rms angle spread of AOA increases, which reduces the correlation between the Tx and Rx elements. (3) Bandwidth effect: MIMO capacity will increase as the signal bandwidth increases. It is because that signal bandwidth becomes larger so that time resolution smaller; hence the antenna array resolves more multipath components, channel which leads to lower correlation among spatial channels, i.e., is higher capacity. (4) Local atterers effect: MIMO capacity increases due to transmitted signals disturbed by local s sc
scatterers so that the angle spread increases, large angle spread leads to high capacity. Thi
Chapter 5 Conclusion
enon can be observed at section 4.5. In indoor environments, when receiver is away out 1.5m, local scatter has no significant effects on MIMO capacity.
eady shown in references [3], [4] and [5]. Our research demonstrates mething new, we also found that (1) MIMO capacity saturates when the spacing is larger than
effects on MIMO capacity.
phenom
from local scatterers ab Some phenomena alr so
one wavelength. (2) When the standard deviation of Tx element spacing is ranging from 1λ to 1.5λ. The capacity of the MIMO with unequal array spacing is always larger than equal spacing. (3) Reference [14] mentions that larger angle spread leads higher capacity, but not to proved it. Our research calculates the angle spread and proves the results. (4) It is also found that signal bandwidth has significant
Reference
Reference
[1] Joseph C. Liberti, Jr.and Theodore S. Rappaport, “Smart Antennas for Wireless Communications: IS-95 and Third Generation CDMA Applications,” Prentice Hall, 1999
[2] G. J. Foschini and M. J. Gans, “On limits of Wireless Communications in a Fading
[2] G. J. Foschini and M. J. Gans, “On limits of Wireless Communications in a Fading