Tải bản đầy đủ - 0 (trang)
2 SS-CDMA: A Proposal for Disaster Message Exchange

2 SS-CDMA: A Proposal for Disaster Message Exchange

Tải bản đầy đủ - 0trang

23 Extended Dependable Air: Use of Satellites in Boosting . . .



679



QZSS



Sho



rt me



ssag

e (C

DMA

Posi

)

tionin

g sig

(Bro

adca

nal

sting

)

Hub station



Terminal 1

Terminal 2



Terminal N



Fig. 23.2 Configuration of SS-CDMA communication system. Positioning signals are broadcast

from QZS and GPS. Each terminal can calculate its own location and time precisely and synchronize its own clock and frequency to those of QZS. The system exhibits high channel gain and can

accommodate several million users using long spread codes and a CDMA scheme



GPS/QZSS

Positioning

Signal



QZSS



Position & Time

Calculation

Short

Message



Timing Control



Spreading



Coding &

Modulation



Message



Freq. Control



Fig. 23.3 Block diagram of SS-CDMA terminal. Each terminal adjusts the transmit timing and

carrier frequency using the location and timing information derived from positioning signals from

satellites (QZS and GPS)



680



K. Tsubouchi et al.



Table 23.1 Channel design parameters of SS-CDMA

Distance of transmission

d = 39,000 km

Carrier frequency

fc = 2.075 GHz

Signal bandwidth

W = 2.24 MHz

Transmission power

A = 30 dBm

Terminal antenna gain

B = −4 dBi

Path loss (free space)

C = 190.6 dB

Fading margin

D = 8 dB

Satellite antenna gain

E = 35 dBi

Spreading gain

Gs

Noise factor (satellite)

F = 5 dB

Temperature (at receiver)

T = 300 K

Required S∕N (BPSK R = 1∕2)

8 dB



S = A + B − C − D + E + Gs [dB]

= 30.0 − 4.0 − 190.6 − 8.0 + 35.0 + Gs [dB]

= −137.6 + Gs [dB]

N = FkTW

= −173.8 + 5.0 + 63.4 = −105.4 [dB]

S∕N = −137.6 + 105.4 + Gs = −32.2 + Gs [dB]



(23.1)

(23.2)

(23.3)



If the required S∕N is 8 dB, the spreading gain Gs must be 40 dB (the spreading

code length is 10,000). The SS-CDMA system aims to achieve multiplexing for 1000

to 10,000 users using a spreading code whose length is 10,000. We adopted the

“accommodation rate” as an efficiency indicator, which is defined as the number

of multiplexing users over a spreading code length. The users assigned to the same

group transmit messages at the same time using a “time slot”. One frame consists of

multiple time slots and contains several million users.

23.2.1.3



Positioning Scheme Using QZSS



Figure 23.4 illustrates the high-accuracy positioning system using QZSS. QZS broadcasts positioning signals of not only L1 coarse/acquisition (L1-C/A), which is one of

the basic GPS signals for civilian use, but also L1 submeter-class augmentation with

integrity function (L1-SAIF), which enhance the positioning performance. In Japan,

the Geospatial Information Authority of Japan (GSI) operates GPS Earth Observation Network (GEONET), which contains over 1200 reference stations to detect the

measurement error of GPS signals at each location. The error is generated by many

factors, such as tropospheric weather conditions, conditions of the ionosphere, and

fluctuation in the satellite orbit. L1-SAIF contains the information used to correct

these errors and is broadcast from QZS via the QZSS control station. The enhanced



23 Extended Dependable Air: Use of Satellites in Boosting . . .



681



GPS satellites

QZSS

Mobile

terminal

L1-SAIF

L1-C/A



GEONET

GSI



Producing

L1-SAIF



QZSS main

control station



ENRI



GEONET: GPS Earth Observation Network

GSI: Geospatial Information Authority of Japan

ENRI: Electronic Navigation Research Institute

Fig. 23.4 High-accuracy positioning system using QZSS. The enhanced GPS technology has been

evaluated using the first QZS, Michibiki, and is expected to achieve submeter positioning accuracy



GPS technology described above has been evaluated using the first QZS, Michibiki,

and is expected to achieve submeter positioning accuracy.



23.2.2 Performance Evaluation by Computer Simulation

In SS-CDMA, all terminals synchronize with each other in the time and frequency

domains using the positioning signals from QZS. The synchronization procedure is

as follows.

1. Each terminal calculates its own location, present time, and reference frequency.

The local clock and phase locked loop (PLL) are synchronized to those of QZS.

2. The terminal precisely calculates its distance from the QZS used as a base station. Each terminal determines the propagation delay using the distance and

related environmental information provided in L1-SAIF.

3. All terminals transmit a modulated signal at a suitable time so that all signals

arrive at the satellite at the same time.

Because not only the chip1 timing but also the Nyquist point is adjusted to the specific timing, as shown in Fig. 23.5, we can minimize the impact of imperfect orthogonality.

In the rest of this subsection, we will show by computer simulation that satisfactory performance in terms of BER can be obtained even in the presence of practical timing jitter and frequency offset if a long-enough code is used in the pro1



The “chip” is a pulse of a direct-sequence spread spectrum (DSSS) code. Chip is also used as units

of period (code length) of SS code.



682



K. Tsubouchi et al.

Transmission Timing

(Terminals)



Reception Timing

(Satellite)



Terminal-1

Arrival target timing

Timing adjustment

Terminal-1

Terminal-2

Terminal-3



Terminal-1

Terminal-2



Terminal-2

Timing adjustment



Terminal-3

Timing jitter



Terminal-3

Timing adjustment



Time



Time



Fig. 23.5 Transmission timing control. Because not only the chip timing but also the Nyquist point

is adjusted to the specific timing, we can minimize the impact of imperfect orthogonality

Table 23.2 Simulation parameters

Modulation



Forward error correction (FEC)

Chip rate

User data rate

Spreading code

Spreading gain

Number of users

Transmitter and receiver filters

Detection

Channel



Binary phase-shift keying (BPSK) (single

carrier), Quadrature phase-shift keying (QPSK)

(single carrier)

Convolutional code R = 1/2

224 kchip/s

220 bit/s

Single M-sequence (rotational shift for each

user)

Gs = 1024

500, 1000

Root cosine roll-off filter

Coherent detection (ideal channel estimation)

Additive white Gaussian noise (AWGN)



posed SS-CDMA method using QZSS. Table 23.2 shows the simulation parameters. The required code gain is 10,000 from the viewpoint of the channel design, but

we adopted one-tenth scaling (Gs = 1024, chip rate is 224 kchip/s) for convenience

of the simulation time. The SS code is a single M-sequence, and rotational shift

sequences are assigned to each user. Figure 23.6 illustrates the code assignment. If

the code length is Gs = 1024 and the number of multiplexing users is Nu , the minimum rotational shift among users is ⌊Gs ∕Nu ⌋, where ⌊x⌋ denotes the maximum

integer that does is not exceed x. In the case of 500 users, the number of shifts is

two, and in the case of 1000 users, it is one. Each terminal uses the rotational shift

code with “0” added last; these codes are referred to as the orthogonal M-sequence.

From the property of the orthogonal M-sequence, the received signal power after the

despreading process is Gs for the same code and 0 for other codes. The orthogonal

M-sequence provides the spreading gain and interference suppression effects.



23 Extended Dependable Air: Use of Satellites in Boosting . . .

Starting position

of SS code

user 2

user 4

user 0

user 1

user 3

user 5



683



user

Gs-2



M-Sequence

Code #

(a) The case of Nu = 1000

Starting position

of SS code

user 0

user 1



user 2



user

(Gs-2)/2



M-Sequence

Code #

(b) The case of Nu = 500



Fig. 23.6 SS code assignment for each user. From the property of the orthogonal M-sequence,

the received signal power after the despreading process is the spreading gain Gs for the same code

and 0 for other codes. The orthogonal M-sequence provides the spreading gain and the interference

suppression effects



23.2.2.1



BER Performance with Timing Jitter



In this subsection, we will elucidate the influence of the timing jitter in the proposed

SS-CDMA. In the condition with timing jitter, it is important that degradation of

S∕N (Eb ∕N0 ∕Gs in the Fig. 23.7) is as small as possible (for example, less than 1 dB)

for the system design of SS-CDMA.

Figure 23.7 shows the bit error rate (BER) performance of SS-CDMA when timing jitter exists. The horizontal axis represents energy per bit to noise power spectral

density ratio (Eb ∕N0 ) considering the spreading gain Gs . In this figure, Nu is the

number of multiplexing users and Tj is the maximum jitter normalized by the chip

period. The lines with open symbols show the BER with Nu = 500, and those with

closed symbols show the BER with Nu = 1000. The timing jitter 𝜏 assigned to each

terminal is random and 𝜏 is distributed uniformly between −Tj and +Tj . The results

in Fig. 23.7 indicate the following.

∙ When Nu = 500, performance degradation is hardly observed in the case of timing

jitter of less than four-eighths chip and only appears when Tj ≥ 6∕8.

∙ The condition Nu = 1000, for which there is a short distance among the codes,

exhibits performance degradation with a small Tj . The degradation of Eb ∕N0 ∕Gs

at a BER of 10−5 is about 0.5 dB with one-eighth-chip timing jitter and over 4 dB

with two-eighths-chip timing jitter.

From these results, we can conclude that timing jitter does not result in performance degradation as long as the distance of the code shift among each user is kept



684



K. Tsubouchi et al.

0



10



10-1



-2



BER



10



-3



10



10-4



-5



10



-30



BPSK Theory

Nu=1,Tj=0

Nu=500,Tj=1/8

Nu=1000,Tj=1/8

Nu=500,Tj=2/8

Nu=1000,Tj=2/8

Nu=500,Tj=4/8

Nu=1000,Tj=4/8

Nu=500,Tj=6/8

Nu=1000,Tj=6/8

-29



-28



-27



-26



-25



-24



-23



Eb/N0/Gs dB



Fig. 23.7 BER performance with timing jitter (BPSK, R = 1∕2, Gs = 1024). The timing jitter does

not result in performance degradation as long as the distance of the code shift among each user is

kept larger than one chip



larger than one chip. In other words, the timing error of the Nyquist point at the satellite is not a dominant factor and is suppressed by the despreading process using a long

SS code. On the other hand, when the distance is less than one chip, the despreading process generates a large intercode interference and causes severe performance

degradation.

QZSS has submeter positioning accuracy. Thus, the reference clock jitter is

expected to be less than 10 ns. When we adopt the channel design parameters in

Table 23.2, the chip period is 446 ns. Therefore, one-eighth-chip timing jitter will be

feasible.



23.2.2.2



BER Performance with Frequency Offset



In this subsection, we will elucidate the influence of the frequency offset in the proposed SS-CDMA. In the condition with frequency offset, it is important that degradation of S∕N is as small as possible (for example, less than 1 dB) for the system

design of SS-CDMA.

Figure 23.8 illustrates the frequency synchronization block diagram using QZSS.

Each terminal produces a reference frequency, which is synchronized with the clock

on the satellite using the positioning signals. Then, the difference between the localtemperature-compensated crystal oscillator (TCXO) and the reference frequency is

periodically estimated. The transmit automatic frequency control (AFC) block gives

phase rotation to the modulated signals so as to cancel the frequency gap. The



23 Extended Dependable Air: Use of Satellites in Boosting . . .



GEOS



685



QZS+GPS



Digital Signal Processing

Transmit AFC

QMOD.



Data



MOD.



Diff.

detection



Gen. ref.

frequency



Positioning



TCXO

about 1ppm



PLL



Fig. 23.8 Transmit frequency control using QZSS. Each terminal produces a reference frequency,

which is synchronized with the clock on the satellite using the positioning signals

0



10



Ofst=0,Nu=1

Ofst=1/16,Nu=500

Ofst=1/16,Nu=1000

Ofst=1/8,Nu=500

Ofst=1/8,Nu=1000

Ofst=2/8,Nu=500

Ofst=2/8,Nu=1000

Ofst=4/8,Nu=500

Ofst=4/8,Nu=1000



-1



10



-2



BER



10



-3



10



-4



10



-5



10



-30



-29



-28



-27

-26

Eb/N0/Gs dB



-25



-24



-23



Fig. 23.9 BER performance with frequency offset (BPSK, R = 1∕2, Gs = 1024). We can observe

the performance degradation of Eb ∕N0 ∕Gs with an increase in the frequency offset



transmit signal from each terminal contains the residual frequency offset and the

phase noise in the PLL block. In this subsection, the effects of the residual frequency

offset are evaluated.



686



K. Tsubouchi et al.

10



0



Ofst=0,Nu=1

Ofst=1/16,Nu=1000

Ofst=1/8,Nu=1000

Ofst=2/8,Nu=1000

Ofst=4/8,Nu=1000



-1



10



-2



BER



10



-3



10



-4



10



-5



10



-30



-29



-28



-27

-26

Eb/N0/Gs dB



-25



-24



-23



Fig. 23.10 BER performance with frequency offset (QPSK, R = 1∕2, Gs = 1024). QPSK is more

sensitive than BPSK owing to the shorter distance of the modulated signal constellation. To suppress

the performance degradation of Eb ∕N0 ∕Gs to less than 1 dB, the normalized frequency offset must

be one-eighth or less. One-eighth of the normalized frequency corresponds to 28 Hz. If 2.68 GHz

is assumed as the carrier frequency, 10 ppb frequency stability is required, which has been realized

using a GPS-synchronized oscillator



Figures 23.9 and 23.10 show the BER performance with the frequency offset of

BPSK and QPSK signals, respectively, where Nu is the number of multiplexing users

and Ofst is the frequency offset normalized by the symbol period (=Gs chips). From

these results, we can observe the performance degradation of Eb ∕N0 ∕Gs with an

increase in the frequency offset. The frequency offset breaks the orthogonality among

multiplexing users and generates interference. The interference power is proportional

to the number of multiplexing users; thus, the performance with 1000 users is worse

than that with 500 users. The results for BPSK and QPSK show a similar tendency.

However, QPSK is more sensitive than BPSK owing to the shorter distance of the

modulated signal constellation. To suppress the performance degradation to less than

1 dB, the normalized frequency offset must be one-eighth or less.

One-eighth of the normalized frequency corresponds to 28 Hz (224 Hz/8). If

2.68 GHz is assumed as the carrier frequency, 10 ppb frequency stability is required,

which has been realized using a GPS-synchronized oscillator (see Sect. 9.4).

From results in the subsections, the proposed SS-CDMA has a potential for being

used the large capacity disaster message exchange system, even if there are timing

jitter and frequency offset.



23 Extended Dependable Air: Use of Satellites in Boosting . . .



687



23.3 Heterogeneous Wireless System with Network

Selection Scheme Using Positioning Information

The demand for broadband services involving wireless communication systems has

increased and wireless data traffic has been increasing sharply. One of potential

ways to satisfy the demand is to use a heterogeneous wireless system, multiple wireless systems are combined, such as MBWA and WLAN systems. Additionally, the

dense deployment of small cells is effective for capacity expansion. For example, the

number of WLAN access points (APs) has been increasing in recent years. In the

Third-Generation Partnership Project (3GPP), small-cell enhancement (SCE) has

been discussed for the fourth-generation mobile networks [7]. Upon increasing the

number of small cells, multiple network cells will overlap in heterogeneous wireless

systems.

For optimum data traffic offloading, users must select a cell with the highest channel quality. For cell selection, users generally discover neighboring cells by listening on the broadcast channel and measuring the instantaneous value of the signal

strength, such as the received signal strength indicator (RSSI) or the reference signal

received power (RSRP). However, using this scheme, network efficiency decreases in

a dense small-cell environment. The reasons for this are as follows. First, users must

discover a large number of cells and measure the instantaneous signal strength for

each of them. More radio resources are used for cell selection in such an environment.

Next, cell selection errors occur because the instantaneous signal strength fluctuates

owing to fading. Users cannot select the cell with the highest channel quality because

of cell selection errors. Finally, users cannot obtain traffic load information only by

measuring the signal strength. Even if users select the cell with the highest signal

strength among neighboring cells, the cell may be congested and the throughput

may not be the highest. For example, a WLAN uses carrier sense multiple access

with collision avoidance (CSMA/CA) as a multiple access scheme, for which the

channel efficiency is greatly decreased when the number of users connected to the

WLAN is large. Therefore, to improve the network efficiency in a dense small-cell

environment, the cell selection scheme must satisfy the following requirements:

∙ Reduced radio resource utilization for cell selection.

∙ Suppression of cell selection errors.

∙ Provide a method of acquiring cell traffic load information.

We focus on the systems that provide high-accuracy positioning information, such

as QZSS in Japan and the Global Navigation Satellite System (GLONASS) in Russia with GPS. Using these systems, we can obtain submeter-resolution positioning

information. In this section, we propose a cell selection scheme using positioning

information [1, 8–11]. In the proposed scheme, users select a cell using their location and channel quality information. The channel quality information at each location is provided for cell selection. Users can estimate the channel quality information

by measuring the channel quality only at their location. Thus, users do not have to

discover cells for cell selection. Furthermore, this channel quality information con-



688



K. Tsubouchi et al.



sists of the average signal strength and traffic load information. Therefore, users can

suppress the cell selection error and select a cell in response to real-time variations

in the traffic load.



23.3.1 Network Selection Scheme Using Positioning

Information

In this section, we explain the proposed cell selection scheme. As mentioned in the

introduction, the proposed scheme utilizes the user’s location and channel quality

information. Knowing the location of the user, channel quality information for neighboring cells is referenced, and the cell that should be selected by the user is decided.

In this chapter, we call this information the channel quality map. Figure 23.11 shows

an image of a channel quality map. The channel quality map tells users which cell

should be selected at each location. As shown in Fig. 23.11, since mobile terminal

(MT) #A is in area AP #1, the MT #A should select cell AP #1. The channel quality

map uses average signal strengths as the physical channel quality of the cells. In the

handover of a heterogeneous wireless system, the average signal strength is a suitable

cell selection criterion. For example, in a cellular system, a high-speed closed-loop

transmission power control and an adaptive modulation and coding (AMC) are used

for following the fluctuation of the signal strength. However, in a self-distributed

wireless system such as a WLAN, there is no such function. The user’s terminal

cannot cope with instantaneous fluctuations of the signal strength in a WLAN. The

handover in a heterogeneous wireless systems is executed very slowly compared with

the fluctuation speed of the signal strength. Therefore, the average signal strength is

used in the channel quality map as a common cell selection criterion in the proposed

scheme.

Thus, in the proposed scheme, it is necessary to construct a channel quality map

in addition to obtaining the positioning information. Figure 23.12 shows an outline

of the method used to construct the channel quality map. To construct the map, calculation of the average signal strength at each location is necessary. In this section,

the user’s measurement of the instantaneous value is utilized as information. The

heterogeneous control server (HCS) in Fig. 23.12 gathers measurement information

from the terminals of an enormous numbers of users and statistically processes the

gathered information to construct the map. The HCS stores the channel quality map,

and users obtain the map from this server. Hereafter, we explain the method of creating the channel quality map. The map is constructed using the following two steps:

(a) Measurement of the signal strength and location

First, to calculate the average signal strength, users measure the instantaneous

signal strength of each cell. At the same time, users also measure their own

location. These two pieces of information are stored by individual users. Even

in the proposed scheme, users have to measure the signal strengths, but this

measurement is executed by an enormous number of users and is not used for cell



23 Extended Dependable Air: Use of Satellites in Boosting . . .



MBWA

BS



A

WLAN

AP #1

AP #2



689



MBWA BS area

WLAN AP #3 area



D

AP #3

C



MT #A: AP #1 is selected.

MT #B: AP #2 is selected.

MT #C: AP #3 is selected.

MT #D: MBWA BS is selected.



B



Fig. 23.11 Image of channel quality map. AP, BS, and MT are access point (WLAN), base station

(MBWA) and mobile terminal, respectively. The channel quality map tells users which cell should

be selected at each location

(a-1) measurement location

MBWA

(b-2) gathering reports

to construct map

(b-1) measurement report



HCS



navigation

satellite system

ex. QZSS



WLAN AP



(a-2) signal strength measurement

HCS: heterogeneous control server



Fig. 23.12 Steps in construction of channel quality map. The map is constructed using the following two steps: a measurement of the signal strength and location and b gathering measurement

reports and constructing the map



selection. Thus, users can carry out this measurement at any time. It is expected

that the measurement time per user will be reduced in the proposed scheme.

(b) Gathering measurement reports and constructing the map

Next, the information discussed in (a) is gathered to construct the map. In this

process, users transmit this information to the HCS. This transmission has a

low real-time demand because the average signal strength is static unless the

neighboring environment greatly changes. Thus, the timing when users transmit

this information can be flexibly adjusted. On the other hand, the HCS executes



Tài liệu bạn tìm kiếm đã sẵn sàng tải về

2 SS-CDMA: A Proposal for Disaster Message Exchange

Tải bản đầy đủ ngay(0 tr)

×