3 Intercarrier Interference (ICI) Compensation in IEEE 802.15.a
3.6 Computer Simulations
x is the preamble in the frequency domain, and I represents an iteration number with an initial value of I = 0, and H0f =Hf
3) Increase the index number I Æ I + 1 and go back to step 2.
Using the DAIC for channel estimation, the receiver can obtain more accurate channel estimates than the LS method. The proposed channel estimation makes the receiver remove ICI and recover the transmitted data accurately.
3.6 Computer Simulations
Computer simulations are conducted to evaluate the performance of the synchronization, channel estimation and proposed ICI compensation methods in the MB OFDM system. In the simulations, the relationship between SNR and Eb/N0 can be defined as
where Es is the symbol energy, Ts is the symbol duration, B is the system bandwidth, and M is the modulation order. When the system transmit power is normalized to one, the noise power given by σ corresponding to a specific E2 b/N0 can be generated by
2 0 b
N σ = E
In the first simulation, the performance of synchronization and channel estimation are investigated in the MB OFDM system. As shown in Figures 3.15-3.18, the bit error rate (BER) of different modes in the CM4 channel decreases with increase in Eb/N0 except for the 480 Mbps data rate mode. Figure 3.18 depicts that BER of the 480 Mbps data rate mode has an error floor resulting from the influence of ICI, because the mode has no strategy for combating the ICI such as spreading or low coding rate. In the next simulation we will investigate the ICI compensation methods.
In the second simulation, the performances of ICI compensation methods are investigated in the MB OFDM system. The ICI induced by the multipath energy outside the FFT window is depicted and the ICI-to-signal ratio is shown in Figure 3.19. To capture sufficient multipath energy and minimize the impact of ICI, the ZPP duration was chosen to be 60.6 ns by the MB OFDM system according to the CM3 channel. As shown in Figure 3.19, the ZPP duration is sufficient to minimize the impact of ICI for CM1-CM3 channels, but not for the CM4 channel. Figure 3.20 shows the uncoded BER versus number of iterations with DAIC for the CM4 channel;
it depicts that performance of the uncoded BER is clearly improved after a single iteration is done. (Synchronization and channel estimation at the receiver are assumed to be perfect.) However, the performance of the uncoded BER will be improved less and less as the number of iterations increases. According to Figure 3.20, we consider that three iterations are enough for the MB OFDM system in the CM4 channel. Figure 3.21 depicts the BER of the 480 Mbps data rate mode in the CM4 channel as a function of Eb/N0. The BER of the conventional method has an error floor due to the impact of ICI. However, the MMSE equalizer and the DAIC
method can improve the performance of the BER significantly. The curve illustrates that the MMSE equalizer in the CM 4 channel gains 1 dB for a BER of 10−3 compared with the DAIC method.
Now, we compare the LS channel estimation with the DAIC channel estimation.
The MSE of two channel estimation techniques in the CM4 channel are shown in Figure 3.22. It shows that the DAIC channel estimation is more accurate than the LS channel estimation at high Eb/N0. However, the MSE of the DAIC channel estimation is higher than the LS channel estimation at low Eb/N0. This is because that errors are dominated by the noise at low Eb/N0, such that the DAIC channel estimation cannot correctly reconstruct the influence of ICI and thus results in the error propagation. As shown in Figure 3.23, the BER performances of the MMSE equalizer and the DAIC method with the LS channel estimation have an error floor. This is because that the LS channel estimation cannot eliminate the ICI, such that the two methods cannot work correctly without the accurate estimated channel. On the other hand, the BER performances of the MMSE equalizer and the DAIC method with the DAIC channel estimation show significant improvement compared with that obtained by the LS channel estimation.
3.7 Summary
In this chapter, we first introduced the indoor UWB channel model and the receiver architecture of the MB OFDM system. The main difference between UWB channel and general narrow-band channel is the “cluster” phenomenon due to the high resolution resulting from large bandwidth of UWB waveforms. There are four channel models defined for various LOS and NLOS environments. In addition, the conventional OFDM receiver techniques like synchronization and channel estimation are then introduced. It is found that the long delay spread in the CM4 channel will
induce the ICI effect and degrade the system performance.
In MB OFDM systems, the ZPP duration chosen to be 60.6ns is enough to capture sufficient multipath energy and minimize the impact of ICI for the CM1-CM3 channels, but not the CM4 channel. For those modes with time-domain spreading, like 106.7 Mbps and 200 Mbps, the MRC is effective to suppress the effect of ICI. In addition, the MMSE equalizer can deal with the influence of ICI for the modes without any spreading, like the 320 Mbps and 480 Mbps modes. However, the high complexity of the channel matrix inversion computation is not suitable for the MB OFDM system. We thus propose the simplified DAIC method to eliminate ICI. The simulation results show that the proposed method can alleviate the influence of the ICI and improve the BER performance in the CM4 channel. More importantly, the proposed method has lower complexity than the MMSE equalizer. It is found that the LS channel estimation is not suitable for the CM4 channel. Due to that the two OFDM training symbols suffer the same influence by ICI, the LS channel estimation cannot perform reliably under the CM4 channel. This prompts us to develop the DAIC method to estimate the channel impulse response. The simulation results indicate that the DAIC channel estimation effectively eliminates the effect of ICI and is more accurate than the LS channel estimation.
fc complex LP system h t H f
( ) ( )
real BP system h t H f
complex lowpass equivalent model ( )
Figure 3.1: Simulation of passband system in terms of equivalent complex baseband system
1 Impulse response realizations
Time (nS)
Figure 3.2: 100 impulse responses based on the CM3 channel model (NLOS up to 10 m with average RMS delay spread of 15 ns)
0 20 40 60 80 100 120 140 160 180 200
Average Power Decay Profile
Delay (nsec)
Average power (dB)
Figure 3.3: Average power decay profile for the channel model CM3 (NLOS up to 10 m with average RMS delay spread of 15 ns)
Figure 3.4: Block diagram of the MB OFDM receiver
...
Figure 3.5: Block diagram of the cross-correlation packet detection
| |
2 After Packet Detection nˆrn
Symbol Timing Detection
Figure 3.6: Block diagram of the symbol timing estimation
Z-D ( )*
Sum Over Preamble Samples
After Packet Detection
Angle (Sum)
rn z
f ˆ
∆Scale
Figure 3.7: Block diagram of the frequency synchronization estimation
Figure 3.8: PSD plots for the OFDM system using CP prefix
Figure 3.9: PSD plots for the OFDM system using ZPP
Figure 3.10: The ZPP OFDM system pick up larger noise at receiver
Zero
Prefix
Noise
Zero
Prefix
Noise
Xn-1 ZP Xn ZP
Xn-1 ZP Xn ZP
Xn-1 ZP Xn ZP
Previous symbol addition
(ISI generation) current symbol addition
(ICI generation) 1st path
2nd path
3rd path
FFT window
Figure 3.11: Illustration of ISI and ICI due to long delay path
0 0
0
ICI inducing terms ISI inducing terms
0
Figure 3.12: Illustration of ISI (right) and ICI (left) channel matrix shapes
0
50
100
0 50
100 -50 -40 -30 -20 -10 0
subcarrier subcarrier
channel magnitude (dB)
Figure 3.13: Magnitude of ICI matrix in frequency domain
FFT FEQ Xi
FHICIFT Received
Signal
Output Data Temporal
Symbol Decision Y
Figure 3.14: Block diagram of the decision-aided ICI canceller
Figure 3.15: Coded BER as a function of Eb/N0 for 53.3 Mbps data rate of the MB OFDM system in CM1-4 channels with parameters estimation
Figure 3.16: Coded BER as a function of Eb/N0 for 106.7 Mbps data rate of the MB OFDM system in CM1-4 channels with parameters estimation
5 6 7 8 9 1 0 1 1 1 2 1 3
Figure 3.17: Coded BER as a function of Eb/N0 for 200 Mbps data rate of the MB OFDM system in CM1-4 channels with parameters estimation
Figure 3.18: Coded BER as a function of Eb/N0 for 480 Mbps data rate of the MB OFDM system in CM1-4 channels with parameters estimation
0 5 10 15 20 25 30
Figure 3.19: Captured multipath energy as a function of ZPP length for CM1-4 channels
Figure 3.20: Uncoded BER versus Eb/N0 with different iteration number in the CM4 channel
ICI-to-Signal Ratio (dB)Bit Error Rate
Eb/N0 (dB)
Figure 3.21: Coded BER as a function of Eb/N0 for 480 Mbps data rate of the MB OFDM system in the CM4 channel
Figure 3.22: Mean square estimation error of the CM4 channel frequency response
Mean squared error
Figure 3.23: Coded BER versus Eb/N0 for 480 Mbps data rate of the MB OFDM system in the CM4 scenario with estimated channel impulse response
0 5 10 15 20 25 30 35
10-3 10-2 10-1 100
Conventional method (DAIC) DAIC (DAIC)
MMSE equalizer(DAIC) Conventional method (LS) DAIC (LS)
MMSE equalizer (LS)
Bit Error Rate
Eb/N0 (dB)
0 5 10 15 20 25 30 35
10-3 10-2 10-1 100
Conventional method (DAIC) DAIC (DAIC)
MMSE equalizer(DAIC) Conventional method (LS) DAIC (LS)
MMSE equalizer (LS)
Bit Error Rate
Eb/N0 (dB)
Table 3.1: Multipath channel target characteristics and model parameters.
Target Channel
Characteristics CM 1 CM 2 CM 3 CM 4 Mean excess delay (nsec) (τm) 5.05 10.38 14.18
RMS delay (nsec) (τrms) 5.28 8.03 14.28 25
NP10dB 35
NP (85%) 24 36.1 61.54
Model Parameters
Λ (1/nsec) 0.0233 0.4 0.0667 0.0667
λ (1/nsec) 2.5 0.5 2.1 2.1
Γ 7.1 5.5 14.00 24.00
γ 4.3 6.7 7.9 12
σ1 (dB) 3.3941 3.3941 3.3941 3.3941
σ2 (dB) 3.3941 3.3941 3.3941 3.3941
σx (dB) 3 3 3 3
Model Characteristics
Mean excess delay (nsec) (τm) 5.0 9.9 15.9 30.1
RMS delay (nsec) (τrms) 5 8 15 25
NP10dB 12.5 15.3 24.9 41.2
NP (85%) 20.8 33.9 64.7 123.3
Channel energy mean (dB) -0.4 -0.5 0.0 0.3
Channel energy std (dB) 2.9 3.1 3.1 2.7
Chapter 4
Interference Avoidance
Transmission Scheme for IEEE 802.15.3a Multi-band OFDM
Due to the lack of coordination of transmissions among different piconets, the conventional multiple accesses are not suitable for multi-piconet environments. In this chapter, the multiple access techniques for intra- and inter-piconet of the IEEE 802.15.3a MB OFDM system will be introduced first. Then, the collision characteristics of SOP will be introduced in the following section. In addition, the interference avoidance transmission scheme will be described. Finally, the performance simulations are shown in Section 4.5.
4.1 Review of Multiple Access Techniques
Three of the major basic multiple access techniques for wireless communications will be reviewed: first is frequency division multiple access (FDMA), second is time division multiple access (TDMA), third is code division multiple access (CDMA) [31].
In FDMA systems, the frequency-time plane is partitioned into non-overlapping frequency bands. Each of them serves a single user. Every user is therefore equipped
with a transmitter for a given, predetermined, frequency band, and a receiver for each band which can be implemented as a single receiver for the entire range with a bank of bandpass filters for the individual bands. The spectral regions between adjacent channels are called guard bands, which help reduce the interference between channels. FDMA is used exclusively for analog cellular systems, even though FDMA can be used for digital systems in theory. Essentially, FDMA splits the allocated spectrum into many subchannels. The main advantage of FDMA is its simplicity. It doesn’t require any coordination or synchronization among users since each user can use its own frequency band without interference. However, this is also the cause of waste especially when the load is momentarily uneven since when one user is idle his share of bandwidth can’t be used by other users. It should be noted that if the users have uneven long term demands, it is possible to divide the frequency range unevenly, i.e., proportional to the demands. However, FDMA is not flexible when adds a new user to the network requiring equipment modification, such as additional filters, in every other user. The key advantages of FDMA can be summarized as follows:
1. Easy implementation
2. Small intersymbol interference
The disadvantages of FDMA can also be summarized as follows:
1. Low flexibility in channel allocation 2. Small channel capacity in cellular system
In TDMA systems, sharing of the communication resource is accomplished by dividing the frequency-time plane into non-overlapping time slots which are transmitted in periodic bursts. Every user is allowed to transmit freely during the time slot assigned to him, that is, the entire system resources are devoted to that user during the assigned time slot. Time is segmented into intervals called frames. Each frame is further partitioned into user assignable time slots. An integer number of time
slots constitutes a burst time or burst. Guard times are allocated between bursts to prevent overlapping of bursts. Each burst is comprised of a preamble and the message portion. The preamble is the initial portion of a burst used for carrier and clock recovery, station identification, and other housekeeping tasks. The message portion of a burst contains the coded information sequence. In some systems, a training sequence is inserted in the middle of the coded information sequence. The advantage of this scheme is that it can aid the receiver in mitigating the effects of the channel and interference. The disadvantage is that it lowers the frame efficiency; that is, the ratio of the bit available for messages to the total frames length.
The TDMA systems are designed for use in a range of environments and situations, form hand portable using in a downtown office to a mobile user traveling at high speed on the freeway. It also supports a variety of services for the end user, such as voice, data, fax, short message services, and broadcast messages. TDMA offers a flexible air interface, providing high performances in capacity, coverage, mobility, and capability to handle different types of user needs. While TDMA is a good digital system, it is still somewhat inefficient since it has no flexibility for varying data rates and has no accommodations for silence in telephone conservation. TDMA also requires strict signaling and timeslot synchronization. A point worth noting is that both FDMA and TDMA system performances degrade in the presence of the multipath fading. More specifically, due to the high data rates of TDMA systems, the time dispersive channel (a consequence of delay spread phenomenon) causes intersymbol interference (ISI). This is a serious problem in TDMA systems thus requiring adaptive techniques to maintain system performance. The key advantages of TDMA can be summarized as follows:
1. Sharing single carrier frequency with multiple users.
2. Non-continuous transmission makes handoff simpler. It means that mobile assisted handoff is possible.
3. Less stringent power control due to the reduced interuser interference.
4. Slots can be assigned on demand (concatenation and reassignment). It means that bandwidth can be supplied on demand.
The disadvantages of TDMA can also be summarized as follows:
1. High synchronization overhead is needed.
2. Equalization is necessary for high data rates.
3. Power envelop will pulsate. It is caused by interfering with other devices.
4. High frequency/slot allocation complexity is needed.
As is clear from the above simple review, in both FDMA and TDMA techniques the number of channels or time slots is fixed for a given system, and a signal channel is allocated to a single user for the whole period of communications. Having a fixed channel or time slot assignment could guarantee the service quality for real-time and constant-bit-rate voice telephony, the main service at that time. However, fixed channel assignment has displayed its lack of efficiency in utilizing the scarce spectrum, particularly with the number of users increasing. CDMA gives a good solution for the problem of efficiency. In a CDMA system the original narrowband user’s information is spread into a much wider spectrum with a high chip rate.
Because each user uses a different uncorrelated code, it is possible to send multiple user’s information on the same frequency spectrum without serious interference in detecting the desired signal at the receiver as long as the correct spreading code is known to the receiver. The signal from each user will have very low power and be seen by other as background noise. Consequently, as long as the total power of noise and multi-user interference is less than a threshold, the information signal can be recovered by correlating the received spread signal with a synchronized replica of the spreading signal. The key advantages of CDMA can be summarized as follows:
1. The system capacity improvement is mainly contributed by improved coding gain.
2. The average transmitted power is less than the average power typically required
by TDMA.
3. CDMA provides robust operation in fading environments by using a Rake receiver.
The disadvantages of CDMA can also be summarized as follows:
1. Requirement for power control.
2. Requirement for long overhead for synchronization.
4.2 Multiple Access in Multi-band OFDM System
IEEE 802.15.3a defines the piconet as the wireless ad hoc data communications system and it allows a number of independent data devices to communicate with each other [32][33]. The communication range of a piconet is usually confined to the area around person or object which covers about 10 m in all directions and surrounds the person or a device. The piconet consists of several components, as shown in Figure 4.1 [32]. The basic component is the devices. Some device is required to occupy the role of the piconet coordinator (PNC) of the piconet. The PNC provides the timing of the piconet with the beacon. In addition, the PNC controls the QoS requirements, power save modes and access control to the piconet.
The IEEE 802.15.3a MB OFDM system has two type of multiple access.
TDMA is used for devices within one piconet. In addition, different piconets which can operate in the same area are distinguished by the use of different TFICs. In the following, the multiple access of intra- and inter-piconet will be introduced.
4.2.1 Time Division Multiple Access (TDMA) for Intra-Piconet Interference Reduction
According to the Specification of the IEEE 802.15.3, timing in the 802.15.3 piconet is based on the superframe. As shown in Figure 4.2 [32], the superframe is composed of three parts:
1. The beacon: it set the timing allocations and communicates management information for the piconet. The beacon consists of the beacon frame and any announce commands sent by the PNC as a beacon extension.
2. The contention access period (CAP): it communicates commands and/or asynchronous data if it is present in the superframe.
3. The channel time allocation period (CTAP): it is composed of channel time allocations (CTAs), which also include management CTAs (MCTAs).
Commands, isochronous streams and asynchronous data connections are transmitted in the CTAP.
The PNC will determine the length of the CAP and communicate to the devices in the piconet via the beacon. However, the PNC is capable of replace the functionality provided in the CAP with MCTAs. MCTAs are also one type of CTA for communications between the devices and the PNC. The CAP uses carrier sense multiple access with collision avoidance (CSMA/CA) for the medium access. In addition, the CTAP uses a standard TDMA protocol and all devices have specified time windows.
4.2.2 Time-Frequency Interleaving Codes for Inter-Piconet Interference Reduction
For the MB OFDM system, the performance of a piconet in the presence of other piconet is an important design consideration. The SIR will determine the performance and it is given as
sig int
SIR P W
P R
= (4.1)
where Psig is the power of the desired signal, P is the power of the interference, int R is the information data rate, and W is the effective bandwidth of the transmitted signal. The first term in Equation 4.1 is the distance separation between the two piconets, while the second term in the Equation 4.1 is indicated as bandwidth expansion factor and is the processing gain available to suppress the interference. In the MB OFDM system, the effective bandwidth is defined as follows:
B DT
s
N N
W T
= × (4.2)
where N is the number of bands, B NDT is the number of data tones, and T is s the symbol duration. From the above discuss, there are two methods to improve performance in the presence of other piconet interference. The first method is to
where N is the number of bands, B NDT is the number of data tones, and T is s the symbol duration. From the above discuss, there are two methods to improve performance in the presence of other piconet interference. The first method is to