The Code Division Multiple Access CDMA willthen provide access for each user within each frequency band and in each beam.. This is done with an on-boardcode division switch which perform
Trang 1as illustrated in Figure 3.1-A Ground users within each beam access the satellite
by CDMA The satellite is equipped with an on-board switch for routinginter- orintra-beam calls The SS/CDMA network is described in detail in Section 3.2 Similarsatellite systems based on TDMA, called Satellite Switched TDMA (SS/TDMA), arepresented elsewhere [2], [3] and [4]
As in the satellite example, CDMA switchingmay also be used in terrestrialapplications These applications include wireless and cable networks that have CDMA
as their access method An example of such a network, called Base-station SwitchedCDMA (BS/CDMA), is illustrated in Figure 3.1-B The BS/CDMA is comprised
of a CDMA exchange node connected to a number of Radio Distribution Points(RDPs) via distribution lines which carry the CDMA signal The exchange node
in this case provides the switchingcapability for establishingconnectivity betweenthe wireless users This wireless network may be used for fixed or mobile services.Similar systems based on TDMA have also been proposed (see [5] and [6]) Reference[5] presents a wireless TDMA switchingsystem which provides connectivity betweenmobile users in a community of interest, while reference [6] presents another TDMAswitchingsystem for fixed service wireless metropolitan area networks In addition
to wireless applications, CDMA has been proposed for standardization in coax-cablenetworks for providingupstream voice, data and video services (see reference [7]) Inthis case, a switchingCDMA device at the exchange node will provide an efficientmechanism for routingCDMA channels between cable users Such an application iscalled Cable-Switched CDMA (CS/CDMA), and is illustrated in Figure 3.1-C Theabove applications, both satellite and terrestrial, are refered to by the term switchedCDMA (SW/CDMA) networks
Copyright © 2001 John Wiley & Sons Ltd ISBNs: 0-471-49184-5 (Hardback); 0-470-84169-9 (Electronic)
Trang 258 CDMA: ACCESS AND SWITCHING
CDMA Exchange Node
SS/CDMA
A The satellite switched CDMA (SS/CDMA)
B The base station switched CDMA (BS/CDMA)
CDMA Exchange Node
H/E H/E
H/E
H/E
PSTN
Coax-cable Network
C The cable switched CDMA (CS/CDMA)
H/E:Head-End PSTN: Public Switched Telephone Network
RDP: Radio Distribution Point
n
Figure 3.1 Switched CDMA (SW/CDMA) networks
Trang 3In this chapter we focus our attention on the satellite switched CDMA system Wepresent the network architecture, the access method and switchingmechanism, anddescribe the design of its system units We also examine the network operation andcontrol algorithm.
3.2 Satellite Switched CDMA (SS/CDMA)
The service needs for future geostationary satellite systems demand direct two-waycommunication between end satellite users havingUltra Small Aperture Terminals(USAT) (antenna dish 26 in diameter) The requirement for this type of service isthe capability of call routingon-board the satellite That is, the satellite will operatenot only as a repeater, but also as a switchingcenter in space Such services, however,can only become economically feasible if the satellite communication capacity andthroughput is sufficiently high while its service quality is comparable to the quality ofwireline service For this reason the system has to provide higher spectral efficiency,but also more efficient utilization of the available mass and power of the spacecraft.Higher spectral efficiency is achieved by using multibeam satellite antennas which allowresuse of the available spectrum Also, the power needs of the transceiver units can bereduced by introducingnew access and modulation methods operatingat a very lowsignal-to-noise ratio in order to allow the use of USAT Also, higher throughput can beachieved with a demand assignment control mechanism, which allows the distribution
of system functionalities between the satellite and end users
The system proposed to meet the above needs is the Satellite Switched Code DivisionMultiple Access (SS/CDMA) The SS/CDMA resolves both the multiple access andthe satellite switchingproblems The uplink access method is based on CDMA,the downlink on Code Division Multiplexing(CDM) and the on-board switchingoncompatible technology which is also code division (CDS) The system operates withdemand assignment control for both access and switching That is, service bandwidthand switch connections are assigned only upon a user request The SS/CDMA canachieve higher spectral efficiency by allowing frequency reuse, i.e reuse of the availablespectrum in every beam of a multibeam satellite In addition, it provides an efficientswitchingmechanism by establishinga direct end-to-end route with minimal on-boardsignal processing and no on-board buffering The access and switching problems areresolved in one step by the demand assignment control mechanism This approach alsoallows system optimization by using an assignment control algorithm to maximizethroughput and to integrate the traffic of circuit calls and data packets A largepopulation of end users may then access the geostationary satellite which providesthe routingof calls and packets between them The system may offer fixed services forcircuit switched calls (voice, data and video) and packet switched data
A related method based on Time Division Multiple Access (TDMA), called SatelliteSwitched TDMA (SS/TDMA), has been proposed in the past for packet switcheddata services, [2], [3] In SS/TDMA the access method is TDMA and the switching
is based on time multiplexing(TMS) A similar TDMA demand assignment system
is also used in the ACTS satellite for low burst rate traffic [4] The TDMA approach,however, requires frequency reuse of 1/4 or 1/7 (dependingon the beamwidth), whileits switch implementation and algorithm control may be more complex for large switchsizes
Trang 460 CDMA: ACCESS AND SWITCHING
Uplink Downlink
Gateway
PSTN
PSDN
ISL
ISL: Inter-Satellite Links
Figure 3.2 The Satellite Switched CDMA (SS/CDMA)
The SS/CDMA system has been developed for AT&T’s VoiceSpan satellite projectand Ka-band application filling(the VoiceSpan project has not been realized) Inthe followingsection we present the system description, in Section 3.2.2 the satelliteswitchingmechanism, in Section 3.2.3 the description of transmitter and the receiverunits, and in Section 3.2.4 the network operation and control
is shown in Figure 3.2
Transmission Rates and Services
The main objective of the satellite network is to provide services with a directconnection to each subscriber The services offered are both circuit switched and packetswitched The circuit switched services are for voice, video and data, while the packetswitched services are only for data The transmission bit rates, the source bit rates andthe quality of each circuit switched service are shown in Table 3.1 The transmissionrate in each channel type includes a source rate, a subrate, framingbits, and a framequality indicator (CRC) The offered rates for voice services are: 16, 32, and 64 Kbps;
Trang 5Table 3.1 Transmission and source bit rates and the corresponding
services
Channel Source Transmiss Service Required
Multiple Access
The SS/CDMA provides both multiple access and switchingto the multibeam satellite.The multiple access problem is resolved by space, frequency and code division Thespace division multiple access is achieved by multibeam antennas in order to reuse theavailable spectrum in each beam The frequency division multiple access is achieved
by segmenting the available spectrum into frequency bands, each having a convenientsize of 10 MHz (see Figure 3.3) The Code Division Multiple Access (CDMA) willthen provide access for each user within each frequency band and in each beam TheCDMA will spread the user data over the bandwidth of 10 MHz
The satellite also performs the switch function That is, user traffic channels will
be switched from any uplink to any downlink beam This is done with an on-boardcode division switch which performs the switchingof the CDMA codes (identifyingtraffic channels) from any uplink CDMA channel in beam-i to any downlink CDMAchannel in beam-j The SS/CDMA system architecture, shown in the block diagram ofFigure 3.4, is comprised of a satellite and the Customer Premises Equipment (CPE).The CPE contains the Subscriber Unit (SU) and the Terminal Equipment (TE) Each
SU is comprised of the Transceiver Unit (TU) and the Call Control Unit (CCU) The
Trang 662 CDMA: ACCESS AND SWITCHING
10 MHz band
For Traffic Channels only
Access Channel only
Each Uplink Beam - i
i = 1, , 32
10 MHz band
Pilot, SYNC and Paging Channels only For Traffic Channels only
Each Downlink Beam - j
j = 1, , 32
Figure 3.3 Frequency band assignments for the SS/CDMA
TU includes the transmitter units for the Access and the Traffic channels (ACTU andTCTU) on the uplink and the receiver units for Synchronization and Paging (S&PRU)
as well as Traffic channels (TCRU) on the downlink The on-board system architecturehas as its basic functional blocks the Code Division Switch (CDS), the Control Unit(CU) and the receiver and transmitter for the Access (ACRU) and Satellite Broadcastchannels (SBTU)
Common Air Interface
The Common Air Interface (CAI) is defined as the interface between the space andthe earth segments of the system, i.e between the satellite and the subscriber units orgateway offices The CAI provides the Control and the Traffic channels The Controlchannels are: the Access in the uplink, and the Pilot, SYNC and Paging in thedownlink These channels operate on an assigned frequency band (see Figure 3.3) ThePilot and the SYNC provide timingand synchronization to the system while the Accessand Paging channels deliver signaling messages to and from the satellite The Trafficchannels, on the other hand, carry voice, data and signaling information between theend subscriber units The multiple access and modulation of the Traffic Channel isbased on the Spectrally Efficient Code Division Multiple Access (SE-CDMA) schemepresented in Chapter 6 The SE-CDMA provides orthogonal separation of Trafficchannels within each beam, as well as between beams On-board the satellite, theTraffic channels are simply switched from an uplink to a downlink beam without anydata decodingor buffering
Trang 7ACRU: Access Channel Receiver Unit
ACTU: Access Channel Transmitter Unit
CCU: Call Control Unit
CDS: Code Division Switch
CU: Control Unit
CPE: Customer Premises Equipment
SBTU: Satellite Broadcast Transmitter Unit S&PRU: SYNC & Paging Receiver Unit SU: Subscriber Unit
TCRU: Traffic Channel Receiver Unit TCTU: Traffic Channel Transmitter Unit TE: Terminal Equipment
SATELLITE
PILOT CHANNEL SYNC CHANNEL
PA GING CHANNEL
TRAFFIC CHANNEL TRAFFIC
CHANNEL
ACCESS CHANNEL
CU
A C U
S B T U ACTU
Code Division Switch
Code Division Switchingallows the implementation of a nonblockingswitchfabric of low complexity (linear to the size of the switch) without any channeldecoding/encoding or buffering on-board, while it maintains compatibility with theSE-CDMA Common Air Interface (CAI) The proposed switchingsystem consists
of Code Division Switch (CDS) modules Each CDS module routes calls between
N uplink and N downlink beams, where each beam contains of a single frequencyband W (W = 10 MHz) The size of the CDS module then is (N L × NL),where L is the number of Traffic channels in the SE-CDMA band (In a particularimplementation, N = 32 and L ≤ 60.) The basic design idea in a CDS module
is to combine the input port Traffic channels into a bus by spreadingthem with
Trang 864 CDMA: ACCESS AND SWITCHINGthe orthogonal code of their destination port This bus is called a Code DivisionBus (CDB) All Traffic channels in the CDB are orthogonally separated, and can
be routed to the destination output by despreadingwith the orthogonal code of theparticular output port The detailed system architectures of the CDS modules arepresented in Chapter 4 The CDS fabric has been shown to be a nonblockingswitchfabric Also, routingvia the CDS fabric will cause no additional interference to theTraffic channels other than the interference introduced at the input satellite link
A complexity analysis and performance assessment of the CDS is also presented inChapter 4
Demand Assignment Control
The demand assignment process provides access and switching to the Subscriber Unit(SU) in the SS/CDMA system That is, the CDMA frequency band and Traffic channelallocations for circuit or packet switched services are made upon a user request.Message requests and assignments are sent via the signaling control channels (Access
in the uplink and Paging in the downlink), while the information data are transmittedvia the Traffic channels The demand assignment approach allows the establishment of
a direct route between the end SUs via the Code Division Switch (CDS) without anybufferingor header processingon board the satellite It also allows dynamic sharing
of system resources for different services while maximizingthe system throughput Abasic description of the Demand Assignment Control process is the following: each SUinitiates a call by sendinga message request to the on-board Control Unit (CU) via theAccess channel The CU will assign (if available) a Traffic channel for the duration ofthe call by allocatinguplink–downlink frequency bands and CDMA codes identifyingthe Traffic channel The CU will then send the assigned Traffic channel information
to the end SUs via the Paging channels, while the switch makes the appropriateconnection for it The end-SU will then begin transmitting on this channel A detaileddescription of this process is given in Section 3.2.4
As described above, the switchingsystem consists of CDS modules Each CDSmodule performs intra-band switchingby routingthe traffic between beams within asingle pair of uplink and downlink frequency bands There is a number of uplink–downlink pairs of frequency bands allocated for Traffic channels (see Figure 3.3),and an equal number of CDS modules correspondingto these pairs The demandassignment algorithm will also be used to handle the inter-module or inter-bandroutingof traffic This is done by the followingprocedure:
upon the arrival of a call, the SU sends a message request via the Access channel tothe on-board Control Unit which assigns an uplink–downlink pair of frequency bandsand sends back the assignment data via the Paging channel to the SUs The SUs thentune up on the assigned frequency bands and use the corresponding CDS module toswitch its traffic The frequency bands for the Access and Paging channels are pre-assigned to each SU Also, this approach requires that each SU is capable of tuningits transceivers (TCTU and TCRU) to the assigned RF frequency upon arrival of acall (No frequency band assignment can be made to TCTU and TCRU before anycall request.)
The proposed method of frequency band assignments for inter-module routing avoidsthe need for additional hardware on board the satellite, while providinga balance of the
Trang 9traffic load amongthe available frequency bands The number of CDS modules will beequal to the number of uplink or downlink frequency bands For reliability purposes, aspare module is added for use in case one fails Also, the demand assignment algorithmwill further optimize system performance by extendingthe size of the Traffic channelpool beyond the single frequency band.
In addition, the demand assignment operation is utilized to integrate circuit andpacket switched services, and maximizes the utilization of the available switchingresources The proposed method is based on the Movable Boundary, and is described
as follows
Given a pool of K orthogonal Traffic channels, Kc out of K will be allocated forcircuit switched calls and Kp for packet switched data Then K = Kc+ Kp (Thetotal number of Traffic channels K is K = qL, where q is the number of frequencybands and L is the number of Traffic channels per frequency band.) Any unused circuittraffic channel may be assigned momentarily for packets Traffic channels allocated forpacket services are not assigned for circuits
Let kc be the number of active circuit calls and kp the number of packets intransmission at a given time instant, then the Traffic channel assignment rules will
be based on conditions (a) kc ≤ Kc, and (b) kc+ kp ≤ Kc Condition (a) indicatesthat no more than Kc circuit calls may be routed to any uplink beam i and downlinkbeam j Similarly, condition (b) indicates that the total number of circuits and packetsadmitted in the uplink beam-i and downlink beam-j, respectively, cannot exceed thebeam capacity K If condition (a) does not hold true after the arrival of any newcircuit call, the call will be blocked Similarly, if condition (b) does not hold true afterthe arrival of a new data packet, the packet will remain buffered in the SU Givenconditions (a) and (b), scheduling algorithms have been designed to maximize theswitch throughput (see Chapter 5)
Array of Parallel ACDCs
Channel
Channel
Channel
" -parallel Data Receivers
1 2 BBF
Trang 1066 CDMA: ACCESS AND SWITCHING
9.8304 Mc/s
BBF 38.4 ks/s
38.4 ks/s
Figure 3.6 The satellite broadcast transmitter unit
3.2.3 Transmitter and Receiver Units
Access Channel
The Access channel operates on the assigned uplink frequency band or bands Thebasic structure of the Access Channel Transmitter Unit (ACTU) provides a channelencoder followed by the spreader and a quadrature modulator The channel encoderhas a rate 1/2 and may be convolutional or turbo Data are then spread by a PNcode gi The PN codes gi have a length of L (L = 210− 1) chips The spreadingchip rate is Rc(Rc = 9.8304 Mc/s), and the CDMA channel nominal bandwidth is W(W ≈ 10 MHz)
Transmissions over the Access channel obey the Spread Spectrum Random Access(SSRA) protocol The SSRA protocol assumes that the Access channel transmissionsare Asynchronous or Unslotted Accordingto SSRA protocol, there is a unique PNcode gi(t) assigned to each beam i Since each ACTU may begin its transmissionrandomly at any time instant (continuous time), the phase offset of the PN code
at the receiver i.e gi(t− nTc) On the receiver side there will be a set of parallelAccess Channel Detection Circuits (ACDC) in order to detect and despread thearrived signal at any phase offset Signals that arrive at the receiver with a phaseoffset of more than one chip will be distinguished and received Unsuccessful messagetransmissions will be retransmitted after a random delay, while messages that aresuccessfully received will be acknowledged All responses to the accesses made on
an Access channel will be received on a correspondingPagingchannel A detaileddescription of the SSRA protocol and its throughtput performance is presented inChapter 7
Trang 11The Access channel message has a preamble and an information data field Thepreamble contains no data and is used to aquire the phase offset of its PN-code.Dependingon the number of parallel ACDCs on the receiver, the preample lengthwill vary, but will not exceed τaq (τaq ≤ 5 msec) The Access channel, in addition
of deliveringaccess messages, will also be used for synchronization of the Trafficchannel On board the satellite is the Access Channel Receiver Unit (ACRU), shown
in Figure 3.5 The ACRU consists of a noncoherent demodulator, an array of parallelAccess Channel Detection Circuits (ACDC) and a pool of k data decoders Thearray of parallel ACDCs provide a combination of parallel with serial aquisitioncircuits Each ACDC searches for synchronization of the message by correlatingover a window of w chips Given L chips the length of the PN code gi, and kthe number of ACDCs, the window size will then be, w = L/K (For example,
if L = 1094 chips and P = 16, then w = 64 chips.) The correlation processtakes place duringthe message preamble usingthe serial search (double dwell)approach The design parameters of the serial search circuit (such as the lengths ofthe dwell times and the correspondingthresholds) are determined so that it meets therequirement for the false alarm and detection probabilities This analysis is presented
in Chapter 7
Consideringthe longround trip satellite propagation delay (200 ms), the mainperformance requirement of the Access channel is to provide a high probability ofsuccess at the first transmission attempt The probability of message success depends(a) on the successful PN-code acquisition duringthe message preample, (b) on theprobability of collision, and (c) on the probability of no bit errors in the message afterchannel decoding The design requirement for successful aquisition with a probability
of (1−10−4) or higher is to have the preamble length two standard deviations above themean aquisition time The successful retention of the message (no bit errors) requiresthat the message has an optimum length In Chapter 7 we also provide an estimate ofthe optimum mumber of ACRU receivers given the total number of Traffic channels
in the system
Satellite Broadcast Channels
The satellite broadcast channels have assigned downlink frequency bands of bandwidth
W (W ≈ 10 MHz) in each beam Figure 3.6 shows the basic structure of the SatelliteBroadcast Transmitter Unit (SBTU) Each SBTU transmits one Pilot, one Sync and
a number of Paging channels Each broadcast channel is identified by two orthogonalcodes (Wk for I and Wn for Q components) and a beam PN-code gj The I and
Q components have different orthogonal and PN-codes All channels within a beamare ‘orthogonally’ separated, while the beams are separated only by PN-codes (semi-orthogonal implementation) After spreading, the satellite broadcast channels aredigitally combined, then modulated and filtered The Pilot channel is transmitting
at all times and contains no data Each satellite beam is identified by the Pilot’s
PN sequence The Sync channel transmits system information for synchronizingandreceivinga Pagingchannel or transmittingon an Access channel The Pagingchannel
is used by the satellite for transmittingpaginginformation and for respondingtoAccess channel requests
Trang 1268 CDMA: ACCESS AND SWITCHING
Outer
Encoder
RS(x,y)
InnerEncoderTURBOrate k/n
MPSKSignalSetMapping
M = 2n
a = cos Φi
b = sin Φi
QuadratureModulatorSpreader
1 It is an orthogonal CDMA scheme which utilizes an optimized concatenation
of error correctingcodes and bandwidth efficient modulation The orthogonalcode of length L chips will span over the entire length of a symbol
2 The concatenated codes are: Reed–Solomon RS(x, y) with a rate x/y as theouter code and Turbo with a rate of k/n as the inner code (Turbo codes werefer to a general class of codes that use serial or parallel concatenation ofconvolutional codes linked by an interleaf One such class uses two parallelrecursive systematic convolutional codes linked by an interleaver.) The inputbits, after framing, first enter the Reed–Solomon code, then the Turbo code,and are then spread and modulated usingM-ary Phase Shift Keying(M-PSK) (M = 2n) (see Figure 3.7) The spreading of the orthogonal sequencewill span over the length of the M-ary symbol at the input of the spreader
3 The SE-CDMA provides orthogonal separation of all Traffic channels withinthe CDMA bandwidth W (W ≈ 10 MHz) This is achieved by assigningorthogonal codes to each Traffic channel In addition, orthogonal and/orPN-codes are used for separatingthe satellite beams (beam codes)
4 The SE-CDMA can be implemented as Fully Orthogonal (FO), MostlyOrthogonal (MO) or Semi-Orthogonal (SO) All of these implementationsprovide orthogonal separation of all of the Traffic channels within each beam
In addition, the FO/SE-CDMA provides orthogonal separation of the first