Motivation System Model Two-Symbol Detection Multiple-Symbol Detection Simulation Summary Differential Dual-Hop Relaying over Time-Varying Rayleigh-Fading Channels M. R. Avendi and Ha H. Nguyen Department of Electrical & Computer Engineering University of Saskatchewan Canada June, 2013 1
20
Embed
Differential Dual-Hop Relaying over Time-Varying Rayleigh-Fading Channels
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
MotivationSystem Model
Two-Symbol DetectionMultiple-Symbol Detection
SimulationSummary
Differential Dual-Hop Relaying over Time-Varying
Rayleigh-Fading Channels
M. R. Avendi and Ha H. Nguyen
Department of Electrical & Computer EngineeringUniversity of Saskatchewan
Amplify-and-Forward (AF), (figure taken from reference 1)
Simplicity of relay function in Amplify-and-Forward relaying1A. Nosratinia, T. E. Hunter, A. Hedayat, ”Cooperative communication in wireless networks,”
Channel information requiredChannel estimation: training symbolsChallenges: Estimation of SR channel, Mobility of users
Non-coherent detection
Differential modulations and demodulationsNo channel estimation required3 dB performance loss between coherent and non-coherentdetection in slow-fading channelsFor fast-fading channels there would be higher loss that needsto be examined!
6
MotivationSystem Model
Two-Symbol DetectionMultiple-Symbol Detection
SimulationSummary
Cooperative Communications
Detection
Coherent detection
Channel information requiredChannel estimation: training symbolsChallenges: Estimation of SR channel, Mobility of users
Non-coherent detection
Differential modulations and demodulationsNo channel estimation required3 dB performance loss between coherent and non-coherentdetection in slow-fading channelsFor fast-fading channels there would be higher loss that needsto be examined!
6
MotivationSystem Model
Two-Symbol DetectionMultiple-Symbol Detection
SimulationSummary
Differential Dual-Hop Relaying (D-DH)
Rayleigh flat-fading channels, hi [k] ∼ CN (0, 1), i = 1, 2 attime index k
Auto-correlation between two channel coefficients, n symbolsapart, E{hi [k]h∗i [k + n]} = J0(2πfin)
Transmission process is divided into two phases
h1[k]h2[k]
Source
Relay
Destination
7
MotivationSystem Model
Two-Symbol DetectionMultiple-Symbol Detection
SimulationSummary
Differential Dual-Hop Relaying: Phase I
Information bits convert to M-PSK symbols: v [k] ∈ V,V = {e j2π(m−1)/M , m = 1, . . . ,M}.Differential encoding: s[k] = v [k]s[k − 1], s[0] = 1
h1[k]
Source
Relay
Destination
Received signal at Relay:x [k] =
√P0h1s[k] + w1[k], w1[k] ∼ CN (0,N0)
8
MotivationSystem Model
Two-Symbol DetectionMultiple-Symbol Detection
SimulationSummary
Differential Dual-Hop Relaying: Phase II
Relay multiplies received signal with A and forwards
b1, b2(θ), b3(θ) depend on system parameters and channelsauto-correlation, E1(x) exponential integral function.
Error Floor
lim(P0/N0)→∞
Pb(E ) =1
4π
π∫
−π
g(θ)1− α2
α2q(θ) + 1− α2dθ (4)
13
MotivationSystem Model
Two-Symbol DetectionMultiple-Symbol Detection
SimulationSummary
Multiple-Symbol Detection
Multiple-Symbol Detection
To overcome error floor
Take N received symbols: y = [ y [1], y [2], . . . , y [N] ]t
y = A√
P0diag{s}diag{h2}h1 + w (5)
where s = [ s[1], · · · , s[N] ]t , h2 = [ h2[1], · · · , h2[N] ]t ,h1 = [ h1[1], · · · , h1[N] ]t and w = [ w [1], · · · ,w [N] ]t .
ML detection:
s = arg maxs∈CN
{
Eh2
{1
πNdet{Ry}exp
(
−yHR−1y y
)}}
. (6)
Ry covariance matrix of y, depends on h2
14
MotivationSystem Model
Two-Symbol DetectionMultiple-Symbol Detection
SimulationSummary
Multiple-Symbol Detection
Replace Ry by Ry = Eh2{Ry}
s = arg mins∈CN
{
yHR−1y y
}
= arg mins∈CN
{‖Us‖2
}(7)
where U = (LHdiag{y})∗ and L is obtained by the Choleskydecomposition of C−1 = LLH , C = A2P0Rh + (1 + A2)N0IN .Rh = toeplitz{ϕ1(0)ϕ2(0), . . . , ϕ1(N − 1)ϕ2(N − 1)}.Solve by sphere decoding with low complexity
15
MotivationSystem Model
Two-Symbol DetectionMultiple-Symbol Detection
SimulationSummary
Illustrative Results
Simulation Setup
Correlated channels h1[k], h2[k] ∼ CN (0, 1)
Normalized Doppler frequencies f1, f2
Three simulation cases:
f1 f1 Channels status
Case I .001 .001 both slow-fading
Case II .01 .001 SR fast-fading
Case III .02 .01 both fast-fading
Amplification factor: A =√
P1/(P0 + N0)
Power allocation: P0 = P1
16
MotivationSystem Model
Two-Symbol DetectionMultiple-Symbol Detection
SimulationSummary
Illustrative Results
Illustrative Results
- BER in different cases using DBPSK
10 15 20 25 30 35 40 45 50 55 6010
−5
10−4
10−3
10−2
10−1
100
Simulation, N=2
Analysis, N=2
MSDSD, N=10, Case II
MSDSD, N=10, Case III
P0/N0 (dB)
BER
Case I
Case II
Case III
Error Floor
17
MotivationSystem Model
Two-Symbol DetectionMultiple-Symbol Detection
SimulationSummary
Illustrative Results
Illustrative Results
BER in different cases using DQPSK
10 15 20 25 30 35 40 45 50 55 6010
−5
10−4
10−3
10−2
10−1
100
Simulation, N=2
Analysis, N=2
MSDSD, N=10, Case II
MSDSD, N=10, Case III
P0/N0 (dB)
BER
Case I
Case II
Case III
Error Floor
18
MotivationSystem Model
Two-Symbol DetectionMultiple-Symbol Detection
SimulationSummary
Summary
Differential dual-hop transmission in time-varying channels
Two-symbol non-coherent detection• Channel time-series model• Bit-error-rate analysis• Error floor in fast fading channels