A Comparison of energy efficiency for UWB Modulations Adil - PowerPoint PPT Presentation

A Comparison of energy efficiency for UWB Modulations Adil ELABBOUBI, Fouzia ELBAHHAR, Marc HEDDEBAUT, Yassine ELHILLALI 1

2 Outline Objectives • UWB modulations descriptive • The system model • UWB modulations comparative • Conclusions and perspectives •

3 Objectives (1/2) Green Communication:  The growing demand of data from mobile networks users greater energy consumption and greater 𝐷𝑃 2 rejections.  Researchers presented a concept called “Green communications”.  The key techniques of Green Communication are: Cognitive Radio: improve the spectrum utilization efficiency  and the network transmission performance. Network coding: remove the redundant routes optimize the  throughput the effect of energy and bandwidth saving. Smart Grid: the combination of new communication  techniques, hardware and software optimization to save energy.

4 Objectives (2/2)  This work is a part of the third category of green communications techniques.  The UWB system is chosen:  low energy consumption  low complexity.  Using an Analytical model to compare the energy consumption of modulation techniques.  Comparing the energy efficiency of some commonly used UWB modulations in a multi-path environment. efficient modulation for future applications.

5 UWB modulations descriptive (1/3) PPM (Pulse Position Modulation):  Transmitting a short pulse with a delay in time if the transmitted bit is 1.  The PPM signal: 𝑡 = 𝐹 𝑢 +∞ 𝑂 𝑡 𝑞(𝑢 − 𝑘𝑈 𝑔 − 𝑑 𝑘 𝑈 𝑑 − 𝑒[𝑘 𝑂 𝑡 ]𝜀) 𝑘=−∞  The duration of a pulse: 𝑈 𝑞 and the bandwidth: B = 1 𝑞 . 𝑈 𝑔 = 𝛾𝑈 𝑞 ( 𝛾 > 100 ), the symbol duration:  𝑈 𝑔 𝑈 𝑡 = 𝑂 𝑡 𝑈

6 UWB modulations descriptive(2/3) PAM (Pulse Amplitude Modulation)  Transmitting a short pulse with different level of Amplitude for 0 and 1.  The PAM signal: 𝑡 = 𝐹 𝑢 𝐹 𝑢 +∞ 𝑂 𝑡 , ] is one of the 𝐵 𝑒[𝑘 𝑂 𝑡 ] 𝑞(𝑢 − 𝑘𝑈 𝑔 − 𝑑 𝑘 𝑈 𝑑 ) 𝑂 𝑡 𝐵 𝑒[𝑘 𝑂 𝑡 𝑘=−∞ − 3 and 𝐹 𝑢 = 𝐹 𝑏𝑤 3 possible amplitude, 𝐵 𝑒[𝑘 𝑂 𝑡 ] = 2d 𝑘 𝑂 𝑡  Parameters are defined like the PPM case.

7 UWB modulations descriptive (3/3) PSM (Pulse Shape Modulation):  Transmitting a short pulse with two different waveforms for 0 and 1.  The PSM signal: 𝐹 𝑢 +∞ 𝑂 𝑡 with 𝑄 𝑒[𝑘 𝑂 𝑡 ] is one of the 𝑡 = 𝑄 𝑒[𝑘 𝑂 𝑡 ] (𝑢 − 𝑘𝑈 𝑔 − 𝑑 𝑘 𝑈 𝑑 ) 𝑘=−∞ possible waveforms.  P arameters are similar to the PPM’s ones.

8 The system Model (1/2)  The time duration to transmit N bit: 𝑈 = 𝑈 𝑏𝑑 + 𝑈 𝑡𝑚 + 𝑈 𝑢𝑠  The energy needed to transmit N bit: 𝑏𝑑 + 𝑄 𝑡𝑚 𝑈 𝑡𝑚 + 𝑄 𝑢𝑠 𝑈 𝑢𝑠 ( 𝑄 𝑏𝑑 ≫ 𝑄 𝑡𝑚 ⇒ 𝑄 𝑡𝑚 = 0 ) 𝐹 = 𝑄 𝑏𝑑 𝑈  The total energy consumed to transmit 1 bit : 𝐹 𝑏 = (𝑄 𝑢 +𝑄 𝑑 )𝑈 𝑏𝑑 +𝑄 𝑢𝑠 𝑈 𝑢𝑠 𝑂 𝑄 𝑢 : the transmission power, 𝑄 𝑑 = 𝑄 𝑑𝑢 + 𝑄 𝑑𝑠 : the power of the transceiver circuitry.

9 The system Model (2/2)  For transmission: 𝑄 𝑑𝑢 = 𝑄 𝑔𝑗𝑚𝑢 𝑞𝑕 + 𝑄 𝑏𝑛𝑞 + 𝑄 ξ 𝑏𝑛𝑞 = α𝑄 𝑢 ( α = η − 1 ) 𝑄  For PAM and PPM: 𝐵𝐸𝐷 𝑄 𝑑𝑠 = 𝑄 𝑀𝑂𝐵 + 𝑄 𝑛𝑗𝑦 + 𝑄 𝑗𝑜𝑢 + 𝑄 𝑞𝑕 + 𝑄 𝑔𝑗𝑚𝑠 + 𝑄  For PSM: 𝐵𝐸𝐷 𝑄 𝑑𝑠 = 𝑄 𝑀𝑂𝐵 + 2(𝑄 𝑛𝑗𝑦 + 𝑄 𝑗𝑜𝑢 ) + 𝑄 𝑞𝑕 + 𝑄 𝑔𝑗𝑚𝑠 + 𝑄

10 UWB modulations comparative (1/10) The average BER in a UWB channel for PPM +∞ and PSM: 𝑄 𝑓 = 𝑅 𝑇𝑂𝑆. 𝑦 𝑔 ℎ 𝑦 𝑒𝑦 0 After approximations: 2 𝑜−1 𝑜! (𝑜 2 𝜏 2  In case of non-severe fading: 𝑄 𝑓 ≤ 𝑇𝑂𝑆 𝑜 exp 2 − 𝑜𝜈) ∀ 𝑜 ∈ 𝑂 (Lognormal approximation) 𝑞λ 1 +(1−𝑞)λ 2  In case of severe fading: 𝑄 (coxian 𝑓 ≈ 𝑇𝑂𝑆 approximation)

11 UWB modulations comparative (2/10) 𝛾 𝑂𝛾 𝑔 , so 𝑈 𝐶 . 𝑈 𝑡 thus 𝑈 𝐶 ( 𝑂 𝑡 = 1 )  𝑈 𝑡 = 𝑂 𝑡 𝑈 𝑡 = 𝑏𝑑 = 𝑂𝑈 𝑏𝑑 =  In case of non-severe fading for n=1: 𝜏 2 𝜏 2 1 1 2 − 𝜈) so SNR= 2 − 𝜈) 𝑄 𝑓 = 𝑇𝑂𝑆 exp ( 𝑄 𝑓 exp ( 𝜏 2 𝐹 𝑡 𝐹 𝑢 𝑄 𝑢 𝑈 1 2 − 𝜈)𝐻 𝑒 ( 𝐻 𝑒 = 𝑁 𝑚 𝑒 𝑙 𝐻 1 ) SNR = 𝐻 𝑒 𝑂 0 then 𝑄 𝑢 𝑈 𝑡 𝑂 0 = 𝐻 𝑒 𝑂 0 = 𝑡 = 𝑂 0 𝑄 𝑓 exp ( 𝜏 2 1 𝑄 𝑢 𝑈 𝑏𝑑 = 𝑂 0 2 − 𝜈)𝐻 𝑒 N 𝑄 𝑓 exp (  Total energy consumption: 𝜏 2 𝑄 𝑑,𝑄𝑄𝑁 −𝑄 𝑏𝑛𝑞 𝑈 1 𝑏𝑑  𝐹 𝑏,𝑄𝑄𝑁 = (1 + α)𝑂 0 2 − 𝜈)𝐻 𝑒 + 𝑄 𝑓 exp ( 𝑂 𝜏 2 𝑄 𝑑,𝑄𝑇𝑁 −𝑄 𝑏𝑛𝑞 𝑈 1 𝑏𝑑  𝐹 𝑏,𝑄𝑇𝑁 = (1 + α)𝑂 0 2 − 𝜈)𝐻 𝑒 + 𝑄 𝑓 exp ( 𝑂

12 UWB modulations comparative (3/10)  In case of severe fading: 𝑓 ≈ 𝑞λ 1 +(1−𝑞)λ 2 so 𝑇𝑂𝑆 = 𝑞λ 1 +(1−𝑞)λ 2 𝑄 𝑇𝑂𝑆 𝑄 𝑓 SNR = 𝐹 𝑡 𝐻 𝑒 𝑂 0 = 𝑄 𝑢 𝑈 𝐹 𝑢 𝑡 = 𝑂 0 𝑞λ 1 +(1−𝑞)λ 2 𝑂 0 = 𝐻 𝑒 𝑂 0 then 𝑄 𝑢 𝑈 𝑡 𝐻 𝑒 𝑄 𝑓 𝑞λ 1 +(1−𝑞)λ 2 𝑏𝑑 = 𝑂 0 𝐻 𝑒 𝑂 𝑄 𝑢 𝑈 𝑄 𝑓  Total energy consumption: 𝑞λ 1 +(1−𝑞)λ 2 𝑄 𝑑,𝑄𝑄𝑁 −𝑄 𝑏𝑛𝑞 𝑈  𝐹 𝑏,𝑄𝑄𝑁 = (1 + α) 𝑂 0 𝑏𝑑 𝐻 𝑒 + 𝑂 𝑄 𝑓 𝑞λ 1 +(1−𝑞)λ 2 𝑄 𝑑,𝑄𝑇𝑁 −𝑄 𝑏𝑛𝑞 𝑈  𝐹 𝑏,𝑄𝑇𝑁 = (1 + α) 𝑂 0 𝑏𝑑 𝐻 𝑒 + 𝑂 𝑄 𝑓

13 UWB modulations comparative (4/10) The average BER in a UWB channel for PAM: +∞ 𝑄 𝑓 = 𝑅 2𝑇𝑂𝑆. 𝑦 𝑔 ℎ 𝑦 𝑒𝑦 0 After approximations : 𝑓 ≤ 2 𝑜−1 𝑜! (𝑜 2 𝜏 2  In case of non-severe fading: 𝑄 2 − 𝑜𝜈) (2𝑇𝑂𝑆) 𝑜 exp ∀ 𝑜 ∈ 𝑂 (Lognormal approximation) 𝑓 ≈ 𝑞λ 1 +(1−𝑞)λ 2  In case of severe fading: 𝑄 (coxian 2𝑇𝑂𝑆 approximation)

14 UWB modulations comparative (5/10) 𝛾 𝑂𝛾  we have 𝑈 𝑔 , so 𝑈 𝐶 . 𝑈 𝑡 thus 𝑈 𝐶 𝑡 = 𝑂 𝑡 𝑈 𝑡 = 𝑏𝑑 = 𝑂𝑈 𝑏𝑑 =  In case of non-severe fading for n=1: 𝜏 2 𝜏 2 1 1 2 − 𝜈) Alors SNR= 2 − 𝜈) 𝑄 𝑓 = 2𝑇𝑂𝑆 exp ( 2𝑄 𝑓 exp ( 𝜏 2 𝐹 𝑡 𝐹 𝑢 𝑄 𝑢 𝑈 1 SNR = 𝐻 𝑒 𝑂 0 donc 𝑄 𝑢 𝑈 𝑡 𝑡 = 𝑂 0 2 − 𝜈 𝐻 𝑒 𝑂 0 = 𝐻 𝑒 𝑂 0 = 2𝑄 𝑓 exp 𝜏 2 1 𝑏𝑑 = 𝑂 0 2 − 𝜈)𝐻 𝑒 N 𝑄 𝑢 𝑈 2𝑄 𝑓 exp (  Total energy consumption: ( 𝜏 2 𝑄 𝑑,𝑄𝐵𝑁 −𝑄 𝑏𝑛𝑞 𝑈 1 𝑏𝑑 𝐹 𝑏 = (1 + α)𝑂 0 2 − 𝜈)𝐻 𝑒 + 2𝑄 𝑓 exp 𝑂

15 UWB modulations comparative (6/10)  In case of severe fading: 𝑓 ≈ 𝑞λ 1 +(1−𝑞)λ 2 so 𝑇𝑂𝑆 = 𝑞λ 1 +(1−𝑞)λ 2 𝑄 2𝑇𝑂𝑆 2𝑄 𝑓 SNR = 𝐹 𝑡 𝐻 𝑒 𝑂 0 = 𝑄 𝑢 𝑈 𝐹 𝑢 𝑡 = 𝑂 0 𝑞λ 1 +(1−𝑞)λ 2 𝑂 0 = 𝐻 𝑒 𝑂 0 then 𝑄 𝑢 𝑈 𝑡 𝐻 𝑒 2𝑄 𝑓 𝑞λ 1 +(1−𝑞)λ 2 𝑄 𝑢 𝑈 𝑏𝑑 = 𝑂 0 𝐻 𝑒 𝑂 2𝑄 𝑓  Total energy consumption: 𝑄 𝑑,𝑄𝐵𝑁 −𝑄 𝑏𝑛𝑞 𝑈  𝐹 𝑏,𝑄𝐵𝑁 = (1 + α)𝑂 0 𝑞λ 1 +(1−𝑞)λ 2 𝑏𝑑 𝐻 𝑒 + 2𝑄 𝑓 𝑂

16 UWB modulations comparative (7/10) Simulation parameters: 𝑑 = 4 𝐻𝐼𝑎 η = 0.6 𝑙 = 3.5 𝑔 𝑂 = 10 6 𝐶 = 500 𝑁𝐼𝑎 𝑂 0 = −170 dBm/Hz 𝑞𝑕 = 25.2 𝑛𝑋 𝑄 𝑀𝑂𝐵 = 7.68 𝑛𝑋 𝑄 𝑛𝑗𝑦 = 15 𝑛𝑋 𝑄 𝑄 𝑗𝑜𝑢 = 2.5 𝑛𝑋 𝐵𝐸𝐷 = 7.6 𝑛𝑋 𝑄 𝑊𝐻𝐵 = 12 𝑛𝑋 𝑄 𝑄 𝐹𝐸 = 10.8 𝑛𝑋 𝑄 𝑔𝑗𝑚𝑢 = 𝑄 𝑁 𝑚 = 40 𝑒𝐶 𝑔𝑗𝑚𝑢𝑠 = 2.5 𝑛𝑋 𝑓 = 10 −3 𝐻 1 = 28 𝑒𝐶 𝛾 =500 𝑄 𝜈 = −0.0039 𝜏 = 0.6883 λ 1 = 4.9 λ 2 = 65.44 p=1.0617 d=10

17 UWB modulations comparative (8/10) Non-severe fading Severe fading

18 UWB modulations comparative (9/10) Non-severe fading Severe fading

19 UWB modulations comparative (10/10) Non-severe fading Severe fading ( 𝐻 𝑒 = 𝑁 𝑚 𝑒 𝑙 𝐻 1 ) ( 𝐻 𝑒 = 𝑁 𝑚 𝑒 𝑙 𝐻 1 )

20 Conclusion and perspectives  PAM is a good candidate for green modulation: Energy Effeciency. • Low hardware complexity. • Perspectives  Short term: Testing the model in multi-user environement • finding the energy efficient spreading sequences. •  Long term: Using the results to design an energy efficient railway balise. •

21 Aknowledgment We thank the Railenium test & research for their participation to support this work.

22 Thank you for your attention