1 of 15

Intelligent Ultra-Wide Band Network

Team Members - Sanyam Jain and Amod Deo

Mentor - Kareem Bonna

Advisor - Dr. Predrag Spasojevic

2 of 15

Sanyam Jain�jain.sanyam@rutgers.edu

ECE Master’s Student�

Amod Deo�amod.deo@rutgers.edu

ECE Master’s Student�

Our Team

3 of 15

Problem Statement

To study the basics of Ultra-Wide Band (UWB) wireless communication.
Simulate clock synchronization in UWB wireless sensor network.

4 of 15

Ultra-Wide Band (UWB)

Fast and stable transmission of data, indoors and outdoors.
Not tied to any frequency.
Signal bandwidth (B_f) exceeds the lesser of 500 MHz or 20% of the arithmetic center frequency (FCC Regulation).

Power Spectral Density (PSD) is limited to

–41.25 dBm/MHz.

B_f = 2 ( (f_H - f_L)/(f_H + f_L) )

Applications in Industry, Sports, Smart,

Homes and many more.

5 of 15

Clock and Clock Synchronization

Every individual sensor in a network has its own clock can be represented C(t) = t, t is ideal or reference time.
Clock will drift away from the ideal time C_i(t) = Θ + f.t ( Θ clock offset and f is clock skew ).
Clock Synchronization is procedure for providing a common notion of time across a distributed system. It is crucial for number of fundamental operations performed by WSN.

Clock Model of Sensor Nodes

6 of 15

Berkeley’s Algorithm for Clock Synchronization

7 of 15

Simulation Output for Berkeley’s Algorithm

Output at Master Node

Output at Slave node

8 of 15

Matlab Simulation for basic communication in baseband with delay computation

Block Diagram Transmission and Reception Using Root Raised Cosine Filter with Fixed Delay Computation

9 of 15

Data symbols for different coding sequence . 6 trailing zeros will be added to flush complete data from RRC filters (both at receiver end and transmitter end) and prevent data loss

A Barker code or Barker sequence is a finite sequence of N values of +1 and −1.

They have low autocorrelation properties and are less likely to interfere with other sequence.

PN codes are deterministic codes that mimic randomness properties.

if the current state and the

generating function of the PN code are known, the future state of the code can be predicted.

10 of 15

Impact of SNR on cross correlation and thus on delay computation

Random Sequence [1 1 -1 . . .]

Barker Sequence

PN Sequence

11 of 15

Mean Absolute Error and Root Mean Squared Error (RMSE) for delay Error plotted for different values of SNR (1000 iterations for each SNR value)

12 of 15

Below is SNR requirement for different coding scheme to maintain accuracy of 1ms.

Barker = -8 dB
PN = -8.9 dB
Random seq([1 1 -1 -1..]) = -5.4 dB

Thus from above plot PN sequence stands out most resilient (out of three) for our accuracy level (1ms)

13 of 15

Conclusion from above plots

Delay computation is adversely impacted with deteriorating values of SNR. So depending on channel, SNR should be maintained at higher values.

Choosing right coding scheme also plays important role in delay computation.

We can use cross correlation between transmitted and received signal to compute delay for a short training sample and correct the actual data with computed delay.

14 of 15

References

Molisch, Andreas F. et al. “IEEE 802 . 15 . 4a channel model-final report.” (2004).
Y. Wu, Q. Chaudhari and E. Serpedin, "Clock Synchronization of Wireless Sensor Networks," in IEEE Signal Processing Magazine, vol. 28, no. 1, pp. 124-138, Jan. 2011, doi: 10.1109/MSP.2010.938757.
S. P. Chepuri, R. T. Rajan, G. Leus and A. van der Veen, "Joint Clock Synchronization and Ranging: Asymmetrical Time-Stamping and Passive Listening," in IEEE Signal Processing Letters, vol. 20, no. 1, pp. 51-54, Jan. 2013, doi: 10.1109/LSP.2012.2222371.
B. M. Sadler and R. J. Kozick, "A Survey of Time Delay Estimation Performance Bounds," Fourth IEEE Workshop on Sensor Array and Multichannel Processing, 2006., Waltham, MA, 2006, pp. 282-288, doi: 10.1109/SAM.2006.1706138.
https://www.mathworks.com/help/comm/