Implementation of Underwater Channel Emulator (1804) Aaron DeMaio, - - PowerPoint PPT Presentation

Implementation of Underwater Channel Emulator (1804)

Aaron DeMaio, Michael Stratton, and Nicholas Gorbenko Advisor: Dr. Peter WIllett Sponsor: The MITRE Corporation

Overview

• Channel Emulation
• Underwater Effects

○ Path losses due to spreading and absorption ○ Ambient Noise ○ Multipath Reverberation ○ Doppler Spreading

• Design

○ Basic Design ○ Setup

• Data Accuracy

○ Impulse Response

• Results so far
• Timeline

Channel Emulation

• Software Defined Radio System

○ Models time-varying underwater channel and effects in software and hardware ○ Design Platform ■ Based in C++ ■ Uses 3 USRP X310 Software Defined Radios ○ Benefits ■ Cheaper ■ More reliable ■ More closely models underwater channels ■ Modular Figure 1: USRP X310 Software Defined Radio

Main Underwater Effects

• Four main problems with underwater channel modeling

○ Path losses due to spreading and absorption ○ Ambient noise ○ Reverberation due to multipath ○ Doppler spreading Figure 2: Multipath reverberation

Path Losses due to spreading and absorption

• Path losses modeled using Thorp’s equation
• The equation dictates that the attenuation is related to the frequency of the signal
• Mostly used to obtain the transfer function

Multipath and Doppler Spreading

• Multipath is modeled with tapped delay line
• The tapped delay line also models doppler spreading due to the delays
• Input signal is modulated with a delay and coefficient and the output is a sum of

these signals

• Each sample is input through this delay line to simulate each signal having many

copies that arrive at different times

Basic Design

• One USRP X310 to act as a transmitter
• A second USRP X310 as a transceiver connected to another laptop
• Laptop will process the received signals and add channel effects
• Final USRP X310 will receive signal

The Setup

*CE = Channel emulator

Basic Design (cont.)

• Creating the channel

○ Code in C++ to model the channel effects ■ Ambient noise modeled with Gaussian Curve ■ Path Losses modeled with spherical spreading ■ Reverberation due to multipath modeled with time-varying response ■ Doppler spreading modeled by adding frequency variations

Checking Data Accuracy

• How can the validity of the channel simulation be established?

○ Impulse response ○ Checking against a MATLAB version done by Milica Stojanovic

• Found through an impulse train

○ Vector [1,0,0,0… inputted into delay line

• Output signal is measured to be the impulse

response

Impulse Response

Multipath Response Part 1

• Channel Response to input data

Multipath Response Part 2

• Delay Response

Noise Response

• Effects of SNR of channel on signals

Spring Timeline

Questions?