Reference no: EM132878482

EGH443 Advanced Telecommunications - Queensland University

Your company recently won a bid to design and implement a multi-carrier wireless communi- cation systems to provide wireless broadband services to a regional city in Queensland. Mobile service providers have overlooked this community due to economic considerations. Deployment of standard communications systems are costly and the Queensland government is looking for a cheaper alternative. In this task you will perform following tasks to design and test a viable wireless communication system for the above application.

Data collected from an extensive measurement campaign commissioned by Queensland government can be used to place base station antennas at strategic locations to maximize reliability and coverage. The channel measurement campaign has found that the wireless channel in the intended coverage area has multi-path characteristics, with a number of delayed paths. A set of measured data is available to estimate the path loss exponent.

You can extract the path loss data and the measured channel delay profile parameters by following the instructions in Part 1. Each group will have a unique set of data based on your student IDs.

Preliminary Instructions

Download all files to be used in Assignment 2 from blackboard and place them into one working directory. Please also unzip the contents of assignment2.zip into this directory.

There are 2 files contained within assignment2.zip:

1. initialise.m - used to generate the data used in this assignment.

2. A2GenData.p - called by initialise.m

Workspace Preparation

Open initialise.m with MATLAB and read its instructions. Proceed to insert your group number and the student IDs into the script and generate your data for this assignment. The channel delay profile data will be stored in ‘channel.mat'.

The initialise.m script only needs to be executed once.

Part 1 - Estimate Path Loss Exponent of the Channel

Large scale fading can be modeled as a combination of the path loss and log normal shadowing. The path loss LdB in dB as a function of distance d is calculated by:

LdB = L0,dB + 10n log10 (d/d0) + Xσ

where L0,dB is the path loss in dB obtained at the reference distance of d0 away from the transmitter, n the path loss exponent and Xσ the shadowing component. The shadowing component is normally distributed when the path loss is specified in dB, and has zero mean with a standard deviation of σ also specified in dB.

In this part you will be estimating the value of n and σ from data in following two variables.

• d: This vector corresponds to the distance between transmitter and receiver at which a measurement was taken. The units are in meters.
• LdBShadow: This matrix stores the data captured in a measurement campaign. Each row of the matrix corresponds to a new measurement trial. The units are in dB.

Present the measured path loss data at different distances in a scatter plot. Describe the reasons for variations observed in the measurement at a given distance.

For each of the distances inside d, calculate the average path loss and plot in a separate figure.

Fit a linear model for the measured data. Estimate the path loss exponent ( n) and path loss at the reference distance (d0). The fitlm function may be a helpful MATLAB function.

With a transmitter and receiver antenna gains of 10 dB and 3 dB respectively and a receiver sensitivity of -92 dBm, estimate the minimum transmitted power required if the receiver is 4 km away from the transmitter. In this calculation you must allow for a 20 dB fade margin between the received power and the receiver sensitivity.

For each of the distance inside d, calculate the standard deviation and find an estimate for the standard deviation of Xσ

Part 2 - Design an OFDM System

You are asked to design a fixed wireless access (FWA) system to provide wireless broadband services to regional Australia. Your system should be capable of offering a 150 Mb/sec download speed using a 40 MHz band centered around 3.6 GHz spectrum. Remember you need to optimise the bandwidth usage and choose the smallest possible modulation scheme to offer good bit- error-rate performance at low transmit power.

2.1 Load the channel channel.mat file and extract the variables pvec and tvec, where pvec and tvec are channel time delay vector in nanoseconds and the delayed relative power vector of the channel in dB respectively. Plot and label the multipath impulse response of the channel.

Design your OFDM system using above information. You need to decide following pa- rameters of your system.
- N - Number of sub-carriers.
- T - OFDM Symbol duration
- Tg - Guard interval
- m - Modulation index (smallest possible.) Justify your selections.
NOTE: The guard interval Tg > 10×στ and the Sub-carrier bandwidth ?(f ) ≤ Bc , where
Bc is the coherence bandwidth of the channel.
Estimate the maximum data rate of your system if the maximum allowable modulation index is 10 (1024-QAM). Compare the this data rate with that of a equivalent single carrier system and comment on whether this single carrier system would be appropriate for a fading channel.
Write a MATLAB code to simulate the performance of the OFDM system in AWGN channel. Simulate and plot the bit error rate performance within the bit-error rate (BER) range from 0.5 to 10-5 and compare with the theoretical bit-error rate performance.
You can use built-in MATLAB functions in this step. Some of the useful MATLAB functions would be, fft, ifft, qammod, awgn and biterr. However, do not use built-in MATLAB functions to implement the OFDM system.
Comment on the observed bit-error-rate performances and describe in detail two methods that can be used by a practical communication system to improve the bit error rate at the receiver.

Part 3 - Performance in Fading Channels

You are able to meet the required data-rate of the AWGN channel. Now you need to simulate the performance of above OFDM system in a Rayleigh fading channel with the impulse response given by pvec and tvec.

The impulse response of the fading channel follows a Raleigh distribution. Let's define a complex Gaussian random variable as h = X +i Y , where both X and Y are Gaussian random variables with zero mean and the same variance. The envelope (square root) of this complex Gaussian random variable has a Rayleigh distribution.

Simulate a complex Gaussian random variable, h with variance 1 using the randn com- mand and plot a normalised histogram of its envelope. Compare the normalised histogram against the theoretical Rayleigh distribution, given by:

fA(a) = a2/σ2 exp{- 2σ2}

If Xn(n = 1, 2, . . . N ) is a vector of modulated symbols, the OFDM signal can be generated by:

xk = ifft(Xn), k = 1, 2, . . . N

Assuming the fading channel is stationery during the OFDM symbol duration, we can find the frequency response of the channel using:

Hn = fft(h) n = 1, . . . N

The FFT output at the receiver can be expressed as:

Yn = HnXn + nn

Therefore, the received signal can be recovered using a single tap equalisation:

X˜n = Yn/Hn = Xn + nn/Hn = Xn + n˜n

Now, X˜n can be demodulated to extract the transmitted data. You can assume that full channel information is available at the receiver for the following simulation. In practice, channel needs to be estimated using pilot symbols.

Write MATLAB code to simulate the performance of OFDM system in flat (single path) Rayleigh fading channel within the Eb/N0 range from 0 dB to 30 dB. Simulate and plot the bit error rate performance and compare with the theoretical bit-error rate performance in fading and AWGN channels.

The power delay profile of the multipath fading channel is given as:

h(τ ) = p(τi)δ(t - τi) i = 1 . . . L, where L is the number of paths. For the simulation we need to normalise the power. The normalisation factor α can be estimated as:

α = 1/ΣL p(τi)

Then the variance of each path can be estimated as:

σ(τi) = αp(τi) Finally, each path can be generated as below:

h(τ ) = .σ(τi) × (randn(1,K) + 1i ∗ randn(1,K))

where K is the number of channel samples.

Simulate and plot the performance of the OFDM system in a multipath Rayleigh fading channel. Compare bit error rate performances in AWGN channel, flat fading channel and multi-path fading channel and comment on the observations.

You can load the channel.mat file to get the multipath power delay profile of the channel.

Part 4 - Error Detection and Correction

The uncoded OFDM system implemented in Part 3 is not providing the reliability required by the Queensland government for Eb/N0 values above ≈ 8 dB.

Implement a forward error correction code to improve the reliability of the received data. You can choose any forward error correction method for this section. You can also use built-in MATLAB functions to implement the chosen error control coding method.

Simulate and plot the bit-error rate of the coded OFDM system in the above multi-path fading channel and compare with the bit error rate performance of the uncoded OFDM system in the multipath-fading channel.

Discuss the effects that a multipath fading channel has on bit errors in the system com- pared with a AWGN channel. Does your forward error correction code address these effects?

What is the data rate of this system after taking into account the new forward error correction encoding system?

The Queensland government also wants the proposed wireless system to be used for mobile banking and other financial activities. You have been instructed to update the system with an appropriate cyclic-redundancy check code (CRC) to ensure bit errors can be detected.

This CRC should be implemented to detect whether the decoded data after the forward error correction code contains any errors. The CRC code should have a word length of n = 16 and a generator polynomial of length n - k = 5.

Demonstrate that this CRC code can detect errors in your system.

Estimate the rate at which a user can transmit data when the system is upgraded with the chosen error control coding and CRC scheme.

Attachment:- Advanced Telecommunications.rar