Optimal Receiver Scheme for Transmitted ... - Semantic Scholar
Journal of Communications Vol. 10, No. 3, March 2015
Optimal Receiver Scheme for Transmitted-Reference Ultra-Wideband System in Coal Mine Ming Li School of Computer Science and Technology, China University of Mining &Technology, Xuzhou 221116, China Email: [email protected]
only one or several integral windows to collect multipath energy without channel estimation and correlators. However, it needs large search area for UWB signals which affects energy capture efficiency. And the use of squarer greatly reduces the output signal to noise ratio. Meanwhile, integration length, decision threshold and the synchronization point must be optimized, which increase the complexity of this method. Transmitted-Reference receiver technology [2] uses the correlation between the reference signal and the receiving signal, which simplifies the system complexity and reduces the synchronization precision. But the system BER and the system emission energy of this method is poor. Besides, the system transmission rate is limited by the interval between the reference pulse and data pulse in a frame. And the system performance is affected by the delay line length. Further, roadway environment in coal mine is a special confined space, where the delay spread of UWB signals transmission is larger than it on the ground. Especially in the case of non-line sight propagation, the number of multipath components in roadway is larger because of obstacles scattering. Therefore, rake correlation receiver in the roadway will need a lot of correlators which greatly increase the receiver volume. In addition, energy detection technology has much more complex design, and the BER performance is unstable in this complex environment of coal mine. TR receiver technology [4] does not require complex and accurate channel estimation, which reduces the system complexity and the synchronization accuracy requirement. However, several problems need to be settled, especially when it is used in coal mines. Furthermore, many pieces of high power electrical equipment are concentrated mostly in the roadway. So the electromagnetic noise is severe. Therefore, in this paper, an average transmitted reference receiver scheme based on orthogonal codes was studied for UWB signals receiving in coal mines to solve the system BER and the transmission rate problems. A scheme model is established. The BER performance of this scheme is analyzed. The effectiveness of the proposed method on BER performance and the transmission rate is demonstrated by simulation and comparison with traditional TR in coal mine roadway environment.
Abstract— An average Transmitted-Reference (TR) receiver scheme based on orthogonal codes is proposed for improving the bit error rate (BER) performance and the transmitting rate of Ultra-wideband (UWB) TR receiver systems in coal mine. Orthogonal codes are used for modulating the reference pulses and the data pulses in each frame of every bit. Since the orthogonality of the two codes, the inter-pulse interference (IPI) is greatly reduced. Then the decoding reference pulses and the data pulses are averaged in a frame, which decreases the noise in the reference pulses and the data pulses. Finally, I demonstrate this optimal receiver scheme by the simulation in the coal mine roadway ultra-wideband channel model. Detailed simulation experiments are presented to validate the effectiveness of this proposed method. The experimental results show that this method can not only efficiently reduce the noise interference in receiving signals, but also can provide higher transmission rate, which make UWB signal receiving in coal mines more accurately. Index Terms—Coal mine, Ultra-wideband, Reference receiver, bit error rate
I.
Transmitted-
INTRODUCTION
Ultra-wideband (UWB) technology has recently attracted widespread attention because of potential applications in high-speed short-range communications. It has many advantages like large bandwidth, large capacity, low power consumption, high speed, strong anti-inference capability, low power consumption, and so on, which is especially suitable for frequency unrestricted environments like underground. Of all UWB systems, they use extremely short pulses, so the main task of receiving UWB signals is effective pulses detection and recovery. On the ground, rake coherent receiver technology, energy detection and transmitted reference receiver technology have usually been used for UWB signals receiving. The rake receiver [1] can effectively combine paths with different delays, obtain the path diversity gain, and improve the transmission characteristics. But it needs high precision synchronization circuit, channel estimation module, and many correlators, which greatly increase the complexity of the system implementation. Energy detection [3] uses Manuscript received December 12, 2014; revised March 24, 2015. This work was supported by National Youth Science Foundation of China under Grant No. 51404258. Corresponding author email: [email protected]. doi:10.12720/jcm.10.3.206-212 ©2015 Journal of Communications
206
Journal of Communications Vol. 10, No. 3, March 2015
II. TRADITIONAL TRANSMITTED-REFERENCE RECEIVER TECHNOLOGY
where d i is the transmitted data; T f is the frame period;
The schematic diagram of traditional TransmittedReference (TR) receiver technology is shown as Fig. 1. A pair of pulses is transmitted in each frame. One is the reference pulse without modulation, and another is the data pulse after modulation. In the receiver, a delay apparatus and a correlator are used for receiving data. In the transmitter, the transmitting signal is generated by a delay apparatus and a modulation. And the interval between the reference pulse and the data pulse is less than the coherence time of the channel. In the receiver, the receiving signal is amplified and filtered at first, which reduces noise interference. The receiving signal with certain delay is generated by a delay apparatus. And the delay time is the interval between the reference pulse and the data pulse. Then the delayed receiving signal and the original receiving signal are multiplied. Finally, the information is obtained after the integrator and the decision output.
E p indicates the energy per pulse; each bit data is
u (t ) is the transmitted data pulse, which period is T p ; transmitted by N s frames, so the length of each bit of data is Tb N sT f ; Td represents the interval between the reference pulse and data pulse in each frame, Tp Td . Assuming the maximum channel delay spread is max , in order to avoid inter-pulse interference (IPI), Td max and T f 2 max are required. At the receiving end, after the BPF with the bandwidth W, the output of the i-th bit data is:
ri (t )
N s 1
[u(t iN sT f
jT f ) ,
j 0
di u(t iNsTf jTf Td )] n(t ) .
(2)
where n(t ) is the zero mean Gaussian noise, whose power spectral density is N0 2 . According to Figure 2, the correlation output of the symbol d i is:
Delay Apparatus
Pulse Generation
Modulation
Z (i)
N s 1
iN s T f jT f Td
iN T jT T j 0
Transmitter
N s 1
s f
f
d
ri (t )ri (t Td )dt
0 [di u 2 (t ) di u(t )n(t Td ) j 0
r(t)
A
Filter Amplifier
Correlator
C
Integrator
D
Decision Output
u(t )n(t ) n(t )n(t Td )]dt
E
B
y(t ) n1 (t ) n2 (t ) n3 (t )
Delay Apparatus
where is the integration time, and max usually;
Receiver
y (t ) is the useful signal; n1 (t ) , n2 (t ) , n3 (t ) is the noise. The energy for each section are calculated, which are:
Fig. 1. Traditional TR receiver principle block diagram.
Fig. 2 shows the demodulation process of the traditional TR by using the BPSK modulation. A, B, C, D and E represent the signals corresponding to Fig. 1. And the red curves indicate the reference pulses. The blue curves are the data pulses. 0
1
Transmit ting Data Td
0
Tb 1
Tf 0
(3)
E y(t ) N s E p .
(4)
E n (t ) N s N0 E p / 2 .
(5)
E n (t ) N N E E n (t ) N W N 2 1 2 2
1
s
2 3
s
0
p
/2.
(6)
2 0
/ 2.
(7)
B
Assuming n1 (t ) , n2 (t ) , n3 (t ) are uncorrelated zero mean Gaussian variables, and the users transmit data with equal probability. The energy of each bit is Eb 2 N s E p ,
C
and the system error rate is:
0
1
A
D E
0
1
0
E 2 Z (i ) , Pb Q Var Z (i ) E 2 y (t ) Q E n12 (t ) E n22 (t ) E n32 (t )
1
Fig. 2. Demodulation process of traditional TR receiver in BPSK modulation.
So the transmitting signal of traditional TR is: N s 1
s(t ) [u (t iN sT f jT f ) ,
©2015 Journal of Communications
N N Q 2( 0 ) 2WN s ( 0 ) 2 Eb Eb
i 0 j 0
di u(t iNsT f jT f Td )] .
(1)
207
1 2
.
, (8)
Journal of Communications Vol. 10, No. 3, March 2015
peaks like PPM modulation. And it is also proved the best modulation scheme in the autocorrelation receiver [9]. The traditional TR transmission signal and the improved TR transmitting signal are shown in Fig. 3. The red indicates the reference pulses. And the blue is the data pulses. Moreover, the reference pulses are not modulated, but the data pulses are on the contrary. As for the problems in traditional TR receivers above, an average transmitted reference receiver scheme based on orthogonal codes is proposed. The principle is shown in Fig. 3. A pair of orthogonal codes is used for encoding and decoding the reference pulses and the data pulses in each frame of each bit. Then the interval between reference pulses and data pulses is reduced to one pulse duration. Since the orthogonality of these codes, the IPI is greatly reduced. Then the decoded reference pulses and the data pulses are averaged in a frame, which decreases the noise in reference pulses and data pulses. Finally, the relevant decision between the reference pulses and the data pulses is done for the receiving signal.
where Q( x) (2 )1 2 e x 2 dx . 2
x
III. AVERAGE TRANSMITTED-REFERENCE RECEIVER SCHEME MODEL BASED ON ORTHOGONAL CODES According to the analysis above, traditional TR receiver technology has several shortcomings. Firstly, because the reference pulses in the receiving signal contain some noise, which reduces system BER. Reference [5] pointed out that the system BER 104 required SNR for 17 ~ 19dB, which is worse compared to the Rake receiver system. And as for transmitting the reference pulse in each frame, the system emission energy efficiency is poor. Besides, in order to avoid interpulse interference (IPI), the interval between the reference pulse and data pulse in each frame is settled as Td max , which limits the transmission rate as
Rb 1 (2 N max ) . So as to address these issues, several improved solutions are given in Ref. [6], Ref. [7] and Ref. [8]. ATR technology [6] reduces the noise effect of reference pulses on system performance by averaging the decoding reference pulses. But at the same time, the accuracy and volume of analog delay lines limits the realization and the BER performance of TR receiver. DTR technology [7] greatly improves the system transmission rate without transmitting the reference pulse by using differential encoding principle and auto-correlation techniques. But the use of differential encoding, the current received signal is affected by both the previous and the later data, so the inherent error propagation phenomenon affects the system performance. The orthogonality of FSR technology [8] used the frequency shift between the pulse trains to maintain orthogonal between reference pulses and data pulses in a bit length without an analog delay line. So in the receiver, the reference pulses and the data pulses are separated by this orthogonality. However, when the system data transmission rate is higher, orthogonality is difficult to satisfy. As a result, the system performance decreases. Meanwhile, the coal mine tunnel space is narrow. The power of electrical equipment is greater, and they placed more concentrated. So the electromagnetic noise in the roadway is severe, which challenges TR receiving technology. Therefore, it is necessary to design a TR receiver system with better antinoise performance and faster transmitting rate according to the special circumstances of the coal mine. Thus, in this section, an average TR receiver scheme based on orthogonal codes was proposed for UWB receiver system in coal mine. In traditional TR systems, each data frame is composed of the reference pulse and the data pulse. In this scheme, BPSK modulation is used for data pulse modulation. This is because that the BPSK modulation performance is superior to that of BPPM [10], and there are no spectral
©2015 Journal of Communications
Traditional TR Td
0
1
1
0
1
Tb
Tf Improved TR
0
0
1
Fig. 3. Transmitted signals of two TR receivers in BPSK modulation.
Orthogonal Code
Pulse Generation
Modu lation
Delay Apparatus
Orthogonal Code
Transmitter BPF
1 Ns
1 Ns
Ns -1
a
r(t + iTs + jTf )
1,j
j=0
Delay Apparatus
Correl ator
Integ rator
Decision Output
Ns -1
a
r(t + iTs + jTf )
2,j
j=0
Receiver
Fig. 4. Improved TR principle based on orthogonal coding block diagram.
Fig. 4 shows a block diagram of the transmitter and receiver of an UWB average TR receiver system based on orthogonal codes. As it shows below, in the transmitter, the delay apparatus generates the transmit signal by a combination of data pulses and reference pulses. And the interval between the reference pulse and the data pulse is one pulse duration. Meanwhile, the orthogonal code modulation is respectively done towards the reference pulse and the data pulse. Then two sets of orthogonal coding sequences are mutually orthogonal. For example, when one-bit data is transmitted by two frames, two rows or two columns of the second order Walsh codes can be used to modulate the reference pulses and data pulses. In the receiver, the receiving signal is filtered and amplified at first, which reduces certain noise interference. Then in current bit data, the reference pulse and the data pulse in
208
Journal of Communications Vol. 10, No. 3, March 2015
time; n is the delay time of each multipath component;
each frame are respectively multiplied by their modulation codes, and then they are summed and averaged in this bit, which obtains the corresponding reference sequences and the data sequences. The time interval between the reference pulse and the data pulse is reused to generate a local template version by the reference pulse sequence. Finally, the local template sequence and the data sequence are multiplied. And the final information is obtained after integration and judgment.
n is the number of multipath components. In order to establish a channel model that can be achieved, the multipaths are chosen as 10dB lower than the strongest L
multipath at most, namely: h(t ) n (t n ) . L n 0
represents the number of multipath satisfying the condition. Here, for simplicity, the channel energy is L
normalized without loss of generality as
n2 1 .
n 0
The signal in the receiver is:
IV. SYSTEM PERFORMANCE ANALYSIS
According to Fig. 4, the transmitted signal of this scheme is:
s(t )
N s 1
[a1, j u(t iN sT f
i - j 0
r (t ) n s(t n ) n(t )
The received signal is firstly passed over the band-pass filter with the bandwidth of W, that is:
jT f ) ,
a2, j di u(t iNsTf jTf Tp )]
(9)
i - j 0
L
n 0
spread obtained by transmitted pulses through the channel; n(t ) is the band-limited noise gotten by the channel additive noise through the band pass filter; It is assumed that the channel additive noise is a zero mean Gaussian
spread is max , in order to avoid inter-pulse interference
and
a
2, j
| j 0,1,
| j 0,1,
, N s 1
, N s 1 are two sets of mutually
orthogonal code sequences. And they have the same length, which is equal to the frame number of one bit data. As to avoid IPI in the receiver end, two sets of mutually orthogonal codes are taken, that is: T a1, n a2, m 0, n m T a1, n a2, m N s , n m
(10)
ri (t )
j 0
jT f ) ,
(15)
where the front part is the reference pulses; the latter part is the data pulses. According to Fig. 3, in the i-th bit data the average reference sequence and the average data sequence is:
(11)
rr ,i (t )
1 Ns
rd ,i (t )
1 Ns
N s 1
a1, j r (t iN sT f j 0
N s 1
a2, j r (t iN sT f j 0
jT f )
(16)
jT f )
(17)
According to Eq. (15), that is:
(12)
n 0
rr ,i (t )
n
represents independent Gaussian random variables with zero mean, and its magnitude is decay exponential with ©2015 Journal of Communications
a1, j p(t iN sT f
j 0
where γ indicates lognormal shadow fading;
N s 1
N s 1
where n(t ) is the additive noise; h(t ) represents the channel impulse response. According to UWB channel model [10], the impulse response of slow fading multipath UWB channel can be expressed by the sum of a number of different functions with different amplitudes and different delay, that is:
h(t ) n (t n )
white noise, and its variance is 2 . Then the bilateral power spectrum of the band-limited noise through the filter is N0 2 . The i-th information bit of the received signal can be expressed as:
a2, j di p(t iN sT f jT f Tp ) n(t )
After transmitted through the channel, the transmitting signal can be expressed as:
r (t ) s(t )* h(t ) n(t )
(14)
where p(t ) n u (t n ) denotes multipath delay
between the reference pulse and data pulse in each frame is T p . Besides, assuming the maximum channel delay
1, j
jT f ) ,
a2, j di p(t iN sT f jT f Tp )] n(t ) .
Assuming the unit pulse energy is E p , the interval
a
N s 1
[a1, j p(t iN sT f
r (t )
where d i is the transmitted data; T f is the frame period; u (t ) is the transmitted data pulse, which period is T p ; E p indicates that each bit data is transmitted by N s frames, so the length of each bit of data is Tb N sT f .
(IPI), T f 2 max is required.
(13)
n 0
1 Ns
N s 1
[a1, j a1, j p(t ) di a1, j a2, j p(t Tp ) j 0
a1, j n(t iNsTf jTf )]
209
Journal of Communications Vol. 10, No. 3, March 2015
a1, j n(t iN sT f
jT f ) .
j 0
where R( ) N0W sin c(W )cos(2 fc ) .As for pre-filter,
(18)
W 1 Tp 1 Tf . So when T f , R( ) 0 . Moreover, in Eq. (26), j1 j2 . So ( j2 j1 )T f Tp Tf , that is：
N s 1
1 Ns
rd ,i (t )
N s 1
1 Ns
p(t )
[a2, j a1, j p(t ) di a2, j a2, j p(t Tp )
E N3 0
j 0
a2, j n(t iN sT f jT f )]
di p(t Tp )
Therefore，put it into Eq. (20), that is:
N s 1
1 Ns
a2, j n(t iN sT f j 0
jT f )
E bˆi E Y E p
(19)
E N N E E N W N
Tp bˆi sign( rd (t )rr (t Tp )dt )
E N12 N0 E p / 2 N s .
Tp
Y N1 N2 N3
2 2
(20)
where is the integral time, usually max ; Y is the useful signal, that is, in each bit data, the sum of the correlation between every data pulse and the processing reference pulse; N1 is the correlation results between data pulses and relevant noise contained in the local template of reference pulses in one bit data; N 2 is the correlation results between the local template of reference pulses and relevant noise contained in the data pulses in one bit data; N 3 is the correlation results between noises in the local template of reference pulses and in the data pulses in one bit data. According to Eq. (8), there are:
Y di p 2 (t ) di p 2 (t ) 0
N1 N2
1 Ns
1 Ns
a1, j n(t iN sT f 0 di p(t ) j 0
Tp
T
a2, j n(t iN sT f j 0
(22)
(24)
E N1 E N2 0 p
N s 1
a1, j n(t j1T f j1 0
V.
(25)
Tp )
a2, j n(t j2T f )
j2 0
1 N s2
N s 1 N s 1
iN sT f Tp
a1, j a2, j iN T T j1 0 j2 0
s f
©2015 Journal of Communications
p
BER PERFORMANCE SIMULATION AND CONCLUSION
In this simulation, the second order Gaussian pulse with the 0.5 ns width is used, and its pulse forming factor is 0.287ns. BPSK modulation is used for data pulses. The communication channel uses coal mine roadway ultrawideband channel model [11]. In coal mine roadway,the path loss model of UWB system can be simply described as a function of frequency and distance:
N s 1
2
on the same condition, which is the best compared with traditional TR, DTR, and FTR. In contrast to Eq. (32), the BER performance of the new method is better than that of ATR scheme.
noise, and its variance is 2 . According to the central limit theorem, N1 , N 2 and N 3 can be approximated by three Gaussian random variables, there are:
s f
(32)
1 2 M 1 N 2W N s N 0 2 0 Pb Q ( ) ( ) based M Eb M Eb
Gaussian approximation method is used to calculate the BER of receivers. n(t ) is zero mean Gaussian white
iN sT f Tp
(31)
x
j 0
iN T T
/ 2 N s2 .
where Q( x) (2 )1 2 e x 2 dx .
N s 1
1 N s2
2 0
jT f ) ,
a1, j n(t iN sT f jT f Tp )dt
E N3
(30)
According to Ref. [6], the BER of ATR is
N s 1
p
/ 2Ns .
j 0
1 N3 2 Ns
p
E 2 bˆ i , Pb Q ˆ Var bi E 2 Y , Q E N12 E N 22 E N32 1 2 1 N W N 0 2 ( ) Q ( 0 ) Ns Ep 2 N s2 E p
p(t Tp ) a2, j n(t iN sT f jT f )dt (23)
p
(29)
According to the Gaussian approximation, N1 , N 2 and N 3 are uncorrelated zero mean Gaussian variables. The users send the data with the same probability, so the system error rate is:
N s 1
Tp
T
jT f )dt
0
2 3
(21)
0
N s 1
(28)
Then according to traditional TR model, similar calculations can be obtained:
According to Fig. 4, the decision variable is:
(27)
PL( f , d ) PL( f ) PL(d )
R(( j2 j1 )T f Tp )dt (26)
210
(33)
Journal of Communications Vol. 10, No. 3, March 2015
Fig. 5 shows the BER simulation curves of traditional TR (CTR) and this TR system (OTR) presented in this paper in the coal mine tunnel under different frame numbers. Two frame numbers are set in this simulation, respectively: N s 7 and N s 15 . From the curves, we can conclude that firstly the BER of this new TR receiving scheme in coal mine roadway is significantly better than that of traditional TR receiver. Secondly, the data transmission rate of this system is log 2 N s 1 , which is better than that of Rb N s (2Tp mds )
where f is the system frequency; and d is the distance between the receiver and the transmitter. And in the tunnel, compared to the reference distance d 0 , the average path loss at location
d of UWB signal is defined:
PL(d ) PL(d0 ) 10n log10 ( where
d0 ) S (d ), d d0 d
(34)
n is the path loss factor; PL(d0 ) is the path loss
d 0 ; S (d ) is the shadow fading, which varies with different test sites. S (d ) is considered
at a reference distance
traditional TR system ( Rb 1 (2 N max ) ). This is because two approximately orthogonal code sequences are used for coding the reference sequence and the data sequence, which will provide more log2 Ns 1 bits transmitted by orthogonal sets. Thirdly, IPI in traditional TR systems is significantly reduced because of this orthogonality.
to be zero-mean Gaussian distribution with a standard deviation . PL( f ) can be defined as:
PL( f ) f
(35)
In LOS and NLOS environments of the coal mine tunnel, the frequency factor is 1.1 (LOS) ， 1.4 (NLOS). Besides, as for small-scale fading feature in coal mine tunnel, average additional delay and RMS delay are used for characterization. Table I shows the mean and the standard deviation of RMS delay and average additional delay obtained by the cumulative distribution curve.
0
10
-1
10
BER
-2
10
-3
10
-4
10
TABLE I: AVERAGE VALUES OF RMS DELAY AND MEAN EXCESS DELAY (IN NANOSECOND)
RMS
Condition
RMS
-5
10
m
RMS
RMS
4
OTR Ns=7 OTR Ns=7 CTR Ns=15 CTR Ns=15 6 8
10
12
14
16
18
20
22
24
(Ep/N0)/dB
Fig. 5. BER curves of two TR receivers in tunnels. RMS
LOS(15dB)
11.8
4.4
22.61
3.4
ACKNOWLEDGMENT
LOS(20dB)
23.6
5.14
33.76
5.72
NLOS(15dB)
29.7
8.8
49.42
12.04
NLOS(20dB)
44.38
10.6
58.30
8.46
The authors gratefully acknowledge the financial support for our work by National Youth Science Foundation of China (No. 51404258). REFERENCES
The distance between transmitting and receiving antennas is 2 meters. The noise is additive white Gaussian noise. The bandwidth of band-pass filter is W 1 Tp 2GHz . Assuming the noise of a filter is large
[1]
[2]
enough, so the signal can be completely through without any energy loss. Besides, it is assumed that the sight path exists between the transmitting antenna and the receiving antenna. The maximum time expansion [11] of the channel is max =49ns . Then the integration time is given
[3]
by 50ns in order to avoid interference between frames. The path loss and shadow fading aren’t considered in this simulation, and the channel energy is
[4]
L
normalized as n2 1 . Thus in the simulation, after 100
[5]
n 0
times Monte Carlo experiments for each SNR, and finally BER performance is given by averaging this 100 BER results. ©2015 Journal of Communications
211
D. Z. Mou, “Researches on non-coherent receiver technology for impulse radio ultra-wideband signals,” Ph.D. dissertation, Beijing Jiaotong University, Beijing, 2011. Y. Jin, H. Liu, K. J. Kim and K.S. Kwak, “A Reconfigurable Digital Receiver for Transmitted Reference Pulse Cluster UWB Communications,” IEEE Trans. on Vehicular Technology, vol. 63, pp. 4734~4740, March 2014. Z. M. X, H. Nie, Z. Z.Chen, H. Khani, W. D. Xiang, and L. Yu, “On the nonlinear Teager-Kaiser operator for energy detection based impulse radio UWB receivers,” IEEE Trans. on Wireless Communications, vol. 13, pp. 2955-2965, April 2014. M. E. Tutay and S Gezici, “Optimal and suboptimal receivers for code-multiplexed transmitted-reference ultra-wideband systems,” Wireless Communications and Mobile Computing, vol. 13, pp. 1435-1449, November 2013. W. Shiji, “Research on energy collection and acquisition of IRUWB signals,” Ph.D. dissertation, Hairpin University of Technology, Hairpin, 2006.
Journal of Communications Vol. 10, No. 3, March 2015
Y. Jin and K. S. Kwak, “A Transmitted Reference Pulse Cluster Averaging UWB Receiver,” IEEE Systems Journal, vol. PP, pp. 1932-8184, December 2014. [7] M. Farhang and J. A. Salehi, “Optimum receiver design for transmitted-reference signaling,” IEEE Transaction Communication, vol. 5, pp. 1589-1598, May 2010. [8] M. E. Tutay, S. Gezici, and H. V. Poor, “Performance analysis of code-multiplexed transmitted-reference ultra-wideband systems,” in Proc. IEEE International Conf. Ultra-Wideband, Bologna, 2011, pp. 111-115. [9] M. Farhang and J. A. Salehi, “Optimum receiver design for transmitted-reference signaling,” IEEE Trans. on Communication, vol. 59, pp. 1589-1598, March 2011. [10] G. R. Aiello and G. D. Rogerson, “Ultra-wideband wireless systems,” IEEE Microwave Magazine, vol. 4, pp. 36-47, June 2003.
[11] Y. Rissafi, L. Talbi, and M. Ghaddar, “Experimental characterization of an UWB propagation channel in underground mines,” IEEE Trans. on Antennas and Propagation, vol. 60, pp. 240-246, January 2012.
[6]
©2015 Journal of Communications
Li Ming was born in Anhui Province, China, in 1984. She received the B.S. degree from the Yangzhou University of China, Yangzhou, in 2005 in electrical engineering. She received the Ph.D. degree with the School of Mechanical Electronic & Information Engineering, China University of Mining & Technology, Beijing, in 2011. She worked as a teacher in School of Computer Science and Technology, China University of Mining & Technology, from 2012. Her research interests include communication and information system.
212
Optimal Receiver Scheme for Transmitted-Reference Ultra-Wideband System in Coal Mine Ming Li School of Computer Science and Technology, China University of Mining &Technology, Xuzhou 221116, China Email: [email protected]
only one or several integral windows to collect multipath energy without channel estimation and correlators. However, it needs large search area for UWB signals which affects energy capture efficiency. And the use of squarer greatly reduces the output signal to noise ratio. Meanwhile, integration length, decision threshold and the synchronization point must be optimized, which increase the complexity of this method. Transmitted-Reference receiver technology [2] uses the correlation between the reference signal and the receiving signal, which simplifies the system complexity and reduces the synchronization precision. But the system BER and the system emission energy of this method is poor. Besides, the system transmission rate is limited by the interval between the reference pulse and data pulse in a frame. And the system performance is affected by the delay line length. Further, roadway environment in coal mine is a special confined space, where the delay spread of UWB signals transmission is larger than it on the ground. Especially in the case of non-line sight propagation, the number of multipath components in roadway is larger because of obstacles scattering. Therefore, rake correlation receiver in the roadway will need a lot of correlators which greatly increase the receiver volume. In addition, energy detection technology has much more complex design, and the BER performance is unstable in this complex environment of coal mine. TR receiver technology [4] does not require complex and accurate channel estimation, which reduces the system complexity and the synchronization accuracy requirement. However, several problems need to be settled, especially when it is used in coal mines. Furthermore, many pieces of high power electrical equipment are concentrated mostly in the roadway. So the electromagnetic noise is severe. Therefore, in this paper, an average transmitted reference receiver scheme based on orthogonal codes was studied for UWB signals receiving in coal mines to solve the system BER and the transmission rate problems. A scheme model is established. The BER performance of this scheme is analyzed. The effectiveness of the proposed method on BER performance and the transmission rate is demonstrated by simulation and comparison with traditional TR in coal mine roadway environment.
Abstract— An average Transmitted-Reference (TR) receiver scheme based on orthogonal codes is proposed for improving the bit error rate (BER) performance and the transmitting rate of Ultra-wideband (UWB) TR receiver systems in coal mine. Orthogonal codes are used for modulating the reference pulses and the data pulses in each frame of every bit. Since the orthogonality of the two codes, the inter-pulse interference (IPI) is greatly reduced. Then the decoding reference pulses and the data pulses are averaged in a frame, which decreases the noise in the reference pulses and the data pulses. Finally, I demonstrate this optimal receiver scheme by the simulation in the coal mine roadway ultra-wideband channel model. Detailed simulation experiments are presented to validate the effectiveness of this proposed method. The experimental results show that this method can not only efficiently reduce the noise interference in receiving signals, but also can provide higher transmission rate, which make UWB signal receiving in coal mines more accurately. Index Terms—Coal mine, Ultra-wideband, Reference receiver, bit error rate
I.
Transmitted-
INTRODUCTION
Ultra-wideband (UWB) technology has recently attracted widespread attention because of potential applications in high-speed short-range communications. It has many advantages like large bandwidth, large capacity, low power consumption, high speed, strong anti-inference capability, low power consumption, and so on, which is especially suitable for frequency unrestricted environments like underground. Of all UWB systems, they use extremely short pulses, so the main task of receiving UWB signals is effective pulses detection and recovery. On the ground, rake coherent receiver technology, energy detection and transmitted reference receiver technology have usually been used for UWB signals receiving. The rake receiver [1] can effectively combine paths with different delays, obtain the path diversity gain, and improve the transmission characteristics. But it needs high precision synchronization circuit, channel estimation module, and many correlators, which greatly increase the complexity of the system implementation. Energy detection [3] uses Manuscript received December 12, 2014; revised March 24, 2015. This work was supported by National Youth Science Foundation of China under Grant No. 51404258. Corresponding author email: [email protected]. doi:10.12720/jcm.10.3.206-212 ©2015 Journal of Communications
206
Journal of Communications Vol. 10, No. 3, March 2015
II. TRADITIONAL TRANSMITTED-REFERENCE RECEIVER TECHNOLOGY
where d i is the transmitted data; T f is the frame period;
The schematic diagram of traditional TransmittedReference (TR) receiver technology is shown as Fig. 1. A pair of pulses is transmitted in each frame. One is the reference pulse without modulation, and another is the data pulse after modulation. In the receiver, a delay apparatus and a correlator are used for receiving data. In the transmitter, the transmitting signal is generated by a delay apparatus and a modulation. And the interval between the reference pulse and the data pulse is less than the coherence time of the channel. In the receiver, the receiving signal is amplified and filtered at first, which reduces noise interference. The receiving signal with certain delay is generated by a delay apparatus. And the delay time is the interval between the reference pulse and the data pulse. Then the delayed receiving signal and the original receiving signal are multiplied. Finally, the information is obtained after the integrator and the decision output.
E p indicates the energy per pulse; each bit data is
u (t ) is the transmitted data pulse, which period is T p ; transmitted by N s frames, so the length of each bit of data is Tb N sT f ; Td represents the interval between the reference pulse and data pulse in each frame, Tp Td . Assuming the maximum channel delay spread is max , in order to avoid inter-pulse interference (IPI), Td max and T f 2 max are required. At the receiving end, after the BPF with the bandwidth W, the output of the i-th bit data is:
ri (t )
N s 1
[u(t iN sT f
jT f ) ,
j 0
di u(t iNsTf jTf Td )] n(t ) .
(2)
where n(t ) is the zero mean Gaussian noise, whose power spectral density is N0 2 . According to Figure 2, the correlation output of the symbol d i is:
Delay Apparatus
Pulse Generation
Modulation
Z (i)
N s 1
iN s T f jT f Td
iN T jT T j 0
Transmitter
N s 1
s f
f
d
ri (t )ri (t Td )dt
0 [di u 2 (t ) di u(t )n(t Td ) j 0
r(t)
A
Filter Amplifier
Correlator
C
Integrator
D
Decision Output
u(t )n(t ) n(t )n(t Td )]dt
E
B
y(t ) n1 (t ) n2 (t ) n3 (t )
Delay Apparatus
where is the integration time, and max usually;
Receiver
y (t ) is the useful signal; n1 (t ) , n2 (t ) , n3 (t ) is the noise. The energy for each section are calculated, which are:
Fig. 1. Traditional TR receiver principle block diagram.
Fig. 2 shows the demodulation process of the traditional TR by using the BPSK modulation. A, B, C, D and E represent the signals corresponding to Fig. 1. And the red curves indicate the reference pulses. The blue curves are the data pulses. 0
1
Transmit ting Data Td
0
Tb 1
Tf 0
(3)
E y(t ) N s E p .
(4)
E n (t ) N s N0 E p / 2 .
(5)
E n (t ) N N E E n (t ) N W N 2 1 2 2
1
s
2 3
s
0
p
/2.
(6)
2 0
/ 2.
(7)
B
Assuming n1 (t ) , n2 (t ) , n3 (t ) are uncorrelated zero mean Gaussian variables, and the users transmit data with equal probability. The energy of each bit is Eb 2 N s E p ,
C
and the system error rate is:
0
1
A
D E
0
1
0
E 2 Z (i ) , Pb Q Var Z (i ) E 2 y (t ) Q E n12 (t ) E n22 (t ) E n32 (t )
1
Fig. 2. Demodulation process of traditional TR receiver in BPSK modulation.
So the transmitting signal of traditional TR is: N s 1
s(t ) [u (t iN sT f jT f ) ,
©2015 Journal of Communications
N N Q 2( 0 ) 2WN s ( 0 ) 2 Eb Eb
i 0 j 0
di u(t iNsT f jT f Td )] .
(1)
207
1 2
.
, (8)
Journal of Communications Vol. 10, No. 3, March 2015
peaks like PPM modulation. And it is also proved the best modulation scheme in the autocorrelation receiver [9]. The traditional TR transmission signal and the improved TR transmitting signal are shown in Fig. 3. The red indicates the reference pulses. And the blue is the data pulses. Moreover, the reference pulses are not modulated, but the data pulses are on the contrary. As for the problems in traditional TR receivers above, an average transmitted reference receiver scheme based on orthogonal codes is proposed. The principle is shown in Fig. 3. A pair of orthogonal codes is used for encoding and decoding the reference pulses and the data pulses in each frame of each bit. Then the interval between reference pulses and data pulses is reduced to one pulse duration. Since the orthogonality of these codes, the IPI is greatly reduced. Then the decoded reference pulses and the data pulses are averaged in a frame, which decreases the noise in reference pulses and data pulses. Finally, the relevant decision between the reference pulses and the data pulses is done for the receiving signal.
where Q( x) (2 )1 2 e x 2 dx . 2
x
III. AVERAGE TRANSMITTED-REFERENCE RECEIVER SCHEME MODEL BASED ON ORTHOGONAL CODES According to the analysis above, traditional TR receiver technology has several shortcomings. Firstly, because the reference pulses in the receiving signal contain some noise, which reduces system BER. Reference [5] pointed out that the system BER 104 required SNR for 17 ~ 19dB, which is worse compared to the Rake receiver system. And as for transmitting the reference pulse in each frame, the system emission energy efficiency is poor. Besides, in order to avoid interpulse interference (IPI), the interval between the reference pulse and data pulse in each frame is settled as Td max , which limits the transmission rate as
Rb 1 (2 N max ) . So as to address these issues, several improved solutions are given in Ref. [6], Ref. [7] and Ref. [8]. ATR technology [6] reduces the noise effect of reference pulses on system performance by averaging the decoding reference pulses. But at the same time, the accuracy and volume of analog delay lines limits the realization and the BER performance of TR receiver. DTR technology [7] greatly improves the system transmission rate without transmitting the reference pulse by using differential encoding principle and auto-correlation techniques. But the use of differential encoding, the current received signal is affected by both the previous and the later data, so the inherent error propagation phenomenon affects the system performance. The orthogonality of FSR technology [8] used the frequency shift between the pulse trains to maintain orthogonal between reference pulses and data pulses in a bit length without an analog delay line. So in the receiver, the reference pulses and the data pulses are separated by this orthogonality. However, when the system data transmission rate is higher, orthogonality is difficult to satisfy. As a result, the system performance decreases. Meanwhile, the coal mine tunnel space is narrow. The power of electrical equipment is greater, and they placed more concentrated. So the electromagnetic noise in the roadway is severe, which challenges TR receiving technology. Therefore, it is necessary to design a TR receiver system with better antinoise performance and faster transmitting rate according to the special circumstances of the coal mine. Thus, in this section, an average TR receiver scheme based on orthogonal codes was proposed for UWB receiver system in coal mine. In traditional TR systems, each data frame is composed of the reference pulse and the data pulse. In this scheme, BPSK modulation is used for data pulse modulation. This is because that the BPSK modulation performance is superior to that of BPPM [10], and there are no spectral
©2015 Journal of Communications
Traditional TR Td
0
1
1
0
1
Tb
Tf Improved TR
0
0
1
Fig. 3. Transmitted signals of two TR receivers in BPSK modulation.
Orthogonal Code
Pulse Generation
Modu lation
Delay Apparatus
Orthogonal Code
Transmitter BPF
1 Ns
1 Ns
Ns -1
a
r(t + iTs + jTf )
1,j
j=0
Delay Apparatus
Correl ator
Integ rator
Decision Output
Ns -1
a
r(t + iTs + jTf )
2,j
j=0
Receiver
Fig. 4. Improved TR principle based on orthogonal coding block diagram.
Fig. 4 shows a block diagram of the transmitter and receiver of an UWB average TR receiver system based on orthogonal codes. As it shows below, in the transmitter, the delay apparatus generates the transmit signal by a combination of data pulses and reference pulses. And the interval between the reference pulse and the data pulse is one pulse duration. Meanwhile, the orthogonal code modulation is respectively done towards the reference pulse and the data pulse. Then two sets of orthogonal coding sequences are mutually orthogonal. For example, when one-bit data is transmitted by two frames, two rows or two columns of the second order Walsh codes can be used to modulate the reference pulses and data pulses. In the receiver, the receiving signal is filtered and amplified at first, which reduces certain noise interference. Then in current bit data, the reference pulse and the data pulse in
208
Journal of Communications Vol. 10, No. 3, March 2015
time; n is the delay time of each multipath component;
each frame are respectively multiplied by their modulation codes, and then they are summed and averaged in this bit, which obtains the corresponding reference sequences and the data sequences. The time interval between the reference pulse and the data pulse is reused to generate a local template version by the reference pulse sequence. Finally, the local template sequence and the data sequence are multiplied. And the final information is obtained after integration and judgment.
n is the number of multipath components. In order to establish a channel model that can be achieved, the multipaths are chosen as 10dB lower than the strongest L
multipath at most, namely: h(t ) n (t n ) . L n 0
represents the number of multipath satisfying the condition. Here, for simplicity, the channel energy is L
normalized without loss of generality as
n2 1 .
n 0
The signal in the receiver is:
IV. SYSTEM PERFORMANCE ANALYSIS
According to Fig. 4, the transmitted signal of this scheme is:
s(t )
N s 1
[a1, j u(t iN sT f
i - j 0
r (t ) n s(t n ) n(t )
The received signal is firstly passed over the band-pass filter with the bandwidth of W, that is:
jT f ) ,
a2, j di u(t iNsTf jTf Tp )]
(9)
i - j 0
L
n 0
spread obtained by transmitted pulses through the channel; n(t ) is the band-limited noise gotten by the channel additive noise through the band pass filter; It is assumed that the channel additive noise is a zero mean Gaussian
spread is max , in order to avoid inter-pulse interference
and
a
2, j
| j 0,1,
| j 0,1,
, N s 1
, N s 1 are two sets of mutually
orthogonal code sequences. And they have the same length, which is equal to the frame number of one bit data. As to avoid IPI in the receiver end, two sets of mutually orthogonal codes are taken, that is: T a1, n a2, m 0, n m T a1, n a2, m N s , n m
(10)
ri (t )
j 0
jT f ) ,
(15)
where the front part is the reference pulses; the latter part is the data pulses. According to Fig. 3, in the i-th bit data the average reference sequence and the average data sequence is:
(11)
rr ,i (t )
1 Ns
rd ,i (t )
1 Ns
N s 1
a1, j r (t iN sT f j 0
N s 1
a2, j r (t iN sT f j 0
jT f )
(16)
jT f )
(17)
According to Eq. (15), that is:
(12)
n 0
rr ,i (t )
n
represents independent Gaussian random variables with zero mean, and its magnitude is decay exponential with ©2015 Journal of Communications
a1, j p(t iN sT f
j 0
where γ indicates lognormal shadow fading;
N s 1
N s 1
where n(t ) is the additive noise; h(t ) represents the channel impulse response. According to UWB channel model [10], the impulse response of slow fading multipath UWB channel can be expressed by the sum of a number of different functions with different amplitudes and different delay, that is:
h(t ) n (t n )
white noise, and its variance is 2 . Then the bilateral power spectrum of the band-limited noise through the filter is N0 2 . The i-th information bit of the received signal can be expressed as:
a2, j di p(t iN sT f jT f Tp ) n(t )
After transmitted through the channel, the transmitting signal can be expressed as:
r (t ) s(t )* h(t ) n(t )
(14)
where p(t ) n u (t n ) denotes multipath delay
between the reference pulse and data pulse in each frame is T p . Besides, assuming the maximum channel delay
1, j
jT f ) ,
a2, j di p(t iN sT f jT f Tp )] n(t ) .
Assuming the unit pulse energy is E p , the interval
a
N s 1
[a1, j p(t iN sT f
r (t )
where d i is the transmitted data; T f is the frame period; u (t ) is the transmitted data pulse, which period is T p ; E p indicates that each bit data is transmitted by N s frames, so the length of each bit of data is Tb N sT f .
(IPI), T f 2 max is required.
(13)
n 0
1 Ns
N s 1
[a1, j a1, j p(t ) di a1, j a2, j p(t Tp ) j 0
a1, j n(t iNsTf jTf )]
209
Journal of Communications Vol. 10, No. 3, March 2015
a1, j n(t iN sT f
jT f ) .
j 0
where R( ) N0W sin c(W )cos(2 fc ) .As for pre-filter,
(18)
W 1 Tp 1 Tf . So when T f , R( ) 0 . Moreover, in Eq. (26), j1 j2 . So ( j2 j1 )T f Tp Tf , that is：
N s 1
1 Ns
rd ,i (t )
N s 1
1 Ns
p(t )
[a2, j a1, j p(t ) di a2, j a2, j p(t Tp )
E N3 0
j 0
a2, j n(t iN sT f jT f )]
di p(t Tp )
Therefore，put it into Eq. (20), that is:
N s 1
1 Ns
a2, j n(t iN sT f j 0
jT f )
E bˆi E Y E p
(19)
E N N E E N W N
Tp bˆi sign( rd (t )rr (t Tp )dt )
E N12 N0 E p / 2 N s .
Tp
Y N1 N2 N3
2 2
(20)
where is the integral time, usually max ; Y is the useful signal, that is, in each bit data, the sum of the correlation between every data pulse and the processing reference pulse; N1 is the correlation results between data pulses and relevant noise contained in the local template of reference pulses in one bit data; N 2 is the correlation results between the local template of reference pulses and relevant noise contained in the data pulses in one bit data; N 3 is the correlation results between noises in the local template of reference pulses and in the data pulses in one bit data. According to Eq. (8), there are:
Y di p 2 (t ) di p 2 (t ) 0
N1 N2
1 Ns
1 Ns
a1, j n(t iN sT f 0 di p(t ) j 0
Tp
T
a2, j n(t iN sT f j 0
(22)
(24)
E N1 E N2 0 p
N s 1
a1, j n(t j1T f j1 0
V.
(25)
Tp )
a2, j n(t j2T f )
j2 0
1 N s2
N s 1 N s 1
iN sT f Tp
a1, j a2, j iN T T j1 0 j2 0
s f
©2015 Journal of Communications
p
BER PERFORMANCE SIMULATION AND CONCLUSION
In this simulation, the second order Gaussian pulse with the 0.5 ns width is used, and its pulse forming factor is 0.287ns. BPSK modulation is used for data pulses. The communication channel uses coal mine roadway ultrawideband channel model [11]. In coal mine roadway,the path loss model of UWB system can be simply described as a function of frequency and distance:
N s 1
2
on the same condition, which is the best compared with traditional TR, DTR, and FTR. In contrast to Eq. (32), the BER performance of the new method is better than that of ATR scheme.
noise, and its variance is 2 . According to the central limit theorem, N1 , N 2 and N 3 can be approximated by three Gaussian random variables, there are:
s f
(32)
1 2 M 1 N 2W N s N 0 2 0 Pb Q ( ) ( ) based M Eb M Eb
Gaussian approximation method is used to calculate the BER of receivers. n(t ) is zero mean Gaussian white
iN sT f Tp
(31)
x
j 0
iN T T
/ 2 N s2 .
where Q( x) (2 )1 2 e x 2 dx .
N s 1
1 N s2
2 0
jT f ) ,
a1, j n(t iN sT f jT f Tp )dt
E N3
(30)
According to Ref. [6], the BER of ATR is
N s 1
p
/ 2Ns .
j 0
1 N3 2 Ns
p
E 2 bˆ i , Pb Q ˆ Var bi E 2 Y , Q E N12 E N 22 E N32 1 2 1 N W N 0 2 ( ) Q ( 0 ) Ns Ep 2 N s2 E p
p(t Tp ) a2, j n(t iN sT f jT f )dt (23)
p
(29)
According to the Gaussian approximation, N1 , N 2 and N 3 are uncorrelated zero mean Gaussian variables. The users send the data with the same probability, so the system error rate is:
N s 1
Tp
T
jT f )dt
0
2 3
(21)
0
N s 1
(28)
Then according to traditional TR model, similar calculations can be obtained:
According to Fig. 4, the decision variable is:
(27)
PL( f , d ) PL( f ) PL(d )
R(( j2 j1 )T f Tp )dt (26)
210
(33)
Journal of Communications Vol. 10, No. 3, March 2015
Fig. 5 shows the BER simulation curves of traditional TR (CTR) and this TR system (OTR) presented in this paper in the coal mine tunnel under different frame numbers. Two frame numbers are set in this simulation, respectively: N s 7 and N s 15 . From the curves, we can conclude that firstly the BER of this new TR receiving scheme in coal mine roadway is significantly better than that of traditional TR receiver. Secondly, the data transmission rate of this system is log 2 N s 1 , which is better than that of Rb N s (2Tp mds )
where f is the system frequency; and d is the distance between the receiver and the transmitter. And in the tunnel, compared to the reference distance d 0 , the average path loss at location
d of UWB signal is defined:
PL(d ) PL(d0 ) 10n log10 ( where
d0 ) S (d ), d d0 d
(34)
n is the path loss factor; PL(d0 ) is the path loss
d 0 ; S (d ) is the shadow fading, which varies with different test sites. S (d ) is considered
at a reference distance
traditional TR system ( Rb 1 (2 N max ) ). This is because two approximately orthogonal code sequences are used for coding the reference sequence and the data sequence, which will provide more log2 Ns 1 bits transmitted by orthogonal sets. Thirdly, IPI in traditional TR systems is significantly reduced because of this orthogonality.
to be zero-mean Gaussian distribution with a standard deviation . PL( f ) can be defined as:
PL( f ) f
(35)
In LOS and NLOS environments of the coal mine tunnel, the frequency factor is 1.1 (LOS) ， 1.4 (NLOS). Besides, as for small-scale fading feature in coal mine tunnel, average additional delay and RMS delay are used for characterization. Table I shows the mean and the standard deviation of RMS delay and average additional delay obtained by the cumulative distribution curve.
0
10
-1
10
BER
-2
10
-3
10
-4
10
TABLE I: AVERAGE VALUES OF RMS DELAY AND MEAN EXCESS DELAY (IN NANOSECOND)
RMS
Condition
RMS
-5
10
m
RMS
RMS
4
OTR Ns=7 OTR Ns=7 CTR Ns=15 CTR Ns=15 6 8
10
12
14
16
18
20
22
24
(Ep/N0)/dB
Fig. 5. BER curves of two TR receivers in tunnels. RMS
LOS(15dB)
11.8
4.4
22.61
3.4
ACKNOWLEDGMENT
LOS(20dB)
23.6
5.14
33.76
5.72
NLOS(15dB)
29.7
8.8
49.42
12.04
NLOS(20dB)
44.38
10.6
58.30
8.46
The authors gratefully acknowledge the financial support for our work by National Youth Science Foundation of China (No. 51404258). REFERENCES
The distance between transmitting and receiving antennas is 2 meters. The noise is additive white Gaussian noise. The bandwidth of band-pass filter is W 1 Tp 2GHz . Assuming the noise of a filter is large
[1]
[2]
enough, so the signal can be completely through without any energy loss. Besides, it is assumed that the sight path exists between the transmitting antenna and the receiving antenna. The maximum time expansion [11] of the channel is max =49ns . Then the integration time is given
[3]
by 50ns in order to avoid interference between frames. The path loss and shadow fading aren’t considered in this simulation, and the channel energy is
[4]
L
normalized as n2 1 . Thus in the simulation, after 100
[5]
n 0
times Monte Carlo experiments for each SNR, and finally BER performance is given by averaging this 100 BER results. ©2015 Journal of Communications
211
D. Z. Mou, “Researches on non-coherent receiver technology for impulse radio ultra-wideband signals,” Ph.D. dissertation, Beijing Jiaotong University, Beijing, 2011. Y. Jin, H. Liu, K. J. Kim and K.S. Kwak, “A Reconfigurable Digital Receiver for Transmitted Reference Pulse Cluster UWB Communications,” IEEE Trans. on Vehicular Technology, vol. 63, pp. 4734~4740, March 2014. Z. M. X, H. Nie, Z. Z.Chen, H. Khani, W. D. Xiang, and L. Yu, “On the nonlinear Teager-Kaiser operator for energy detection based impulse radio UWB receivers,” IEEE Trans. on Wireless Communications, vol. 13, pp. 2955-2965, April 2014. M. E. Tutay and S Gezici, “Optimal and suboptimal receivers for code-multiplexed transmitted-reference ultra-wideband systems,” Wireless Communications and Mobile Computing, vol. 13, pp. 1435-1449, November 2013. W. Shiji, “Research on energy collection and acquisition of IRUWB signals,” Ph.D. dissertation, Hairpin University of Technology, Hairpin, 2006.
Journal of Communications Vol. 10, No. 3, March 2015
Y. Jin and K. S. Kwak, “A Transmitted Reference Pulse Cluster Averaging UWB Receiver,” IEEE Systems Journal, vol. PP, pp. 1932-8184, December 2014. [7] M. Farhang and J. A. Salehi, “Optimum receiver design for transmitted-reference signaling,” IEEE Transaction Communication, vol. 5, pp. 1589-1598, May 2010. [8] M. E. Tutay, S. Gezici, and H. V. Poor, “Performance analysis of code-multiplexed transmitted-reference ultra-wideband systems,” in Proc. IEEE International Conf. Ultra-Wideband, Bologna, 2011, pp. 111-115. [9] M. Farhang and J. A. Salehi, “Optimum receiver design for transmitted-reference signaling,” IEEE Trans. on Communication, vol. 59, pp. 1589-1598, March 2011. [10] G. R. Aiello and G. D. Rogerson, “Ultra-wideband wireless systems,” IEEE Microwave Magazine, vol. 4, pp. 36-47, June 2003.
[11] Y. Rissafi, L. Talbi, and M. Ghaddar, “Experimental characterization of an UWB propagation channel in underground mines,” IEEE Trans. on Antennas and Propagation, vol. 60, pp. 240-246, January 2012.
[6]
©2015 Journal of Communications
Li Ming was born in Anhui Province, China, in 1984. She received the B.S. degree from the Yangzhou University of China, Yangzhou, in 2005 in electrical engineering. She received the Ph.D. degree with the School of Mechanical Electronic & Information Engineering, China University of Mining & Technology, Beijing, in 2011. She worked as a teacher in School of Computer Science and Technology, China University of Mining & Technology, from 2012. Her research interests include communication and information system.
212
