C ALIFORNIA I NSTITUTE OF T ECHNOLOGY D EPARTMENT OF E LECTRICAL E NGINEERING
EE 163B Communication Theory II
Spring 2005 http://ee163.caltech.edu Design and Implementation of a Wireless Acoustic Communication System This project aims to design and implement an end-to-end digital wireless acoustic communication system, in which a speaker and a microphone are employed as the transmitting and receiving antennas respectively. You will use Matlabs built-in capabilities that let you emit a sound from the computer speaker, and record a sound incident upon the microphone (e.g. see functions soundsc, sound, wavread, auread, . . .). Many sources of noise and distortion will be present in this system, among them, non-ideal and/or nonlinear speaker and microphone frequency responses, multipath propagation between the speaker and the microphone, propagation delay, and ambient noise. Your goal should be to overcome these challenges and design and implement a system that achieves high bit rate throughput and reliable transmission. 1 Specically, your primary design objective should be the following: Maximize the data bit rate Rb , while maintaining a BER of less than 2%, assuming that the microphone and speaker are separated by 15 cm. You have full control over what to choose for the transmit lter, modulator, receive lter, and equalizer, and you are encouraged to try a number of different combinations of these. It is expected that you will implement the types of equalizers discussed in class and compare their performance. It is also expected that you will make smart choices about the type(s) of modulation to choose, and compare their performances as well. Important points: Do not forget practical necessities such as carrier and symbol synchronization, and automatic gain control. For instance, it will be helpful if you use a training (known) sequence or tone, to help in symbol timing and also with equalization (see the gure, which does not show carrier synchronization or AGC blocks). It is expected that you will generate with eye diagrams for your system, both before and after equalization. Once you meet the above design objective generate a plot of data rate R b vs. distance, for different values of total (average) received energy per bit, i.e., set the speaker volume at low, medium, and high; devise a mechanism for accurately (important!) computing the average received energy per bit for each case; vary the distance for each volume setting, and measure the maximum bit rate achievable. Analyze tradeoffs. You can use one or two computers; and you may work in groups of 2, or by yourself. One person from each group must inform me about the names of team members prior to start. On Thursday June 2, each team will be asked to demo the system on a computer of your choice. You will be asked to transmit an image of your choice across the link (see the class web site for image samples). You are also required to submit a detailed 8-10 page report, outlining strategies used, your design, results, insights, and conclusions, via email only, to me. The report, in compressed postscript, PDF, or Word, should contain an appendix, not included in the page limit of 8-10 pages, consisting of a very well documented version of your computer code. The report - email only - is due by 11:00 pm on Saturday June 4. This is a hard deadline. The project will be graded based on the quality of your design, how well you satisfy or exceed the design objective, and the organization, clarity, and level of detail of your report.
1
To make everything fair for all students, you are required not to use data compression or error correction/detection coding in this system.
random generator speaker train MUX symbol mapper g a
g(t)
bband
Re{ }
transmit
image bits
ak e j2f0t
0 -10 -20 -30 -40 0
microphone
EE 6058 Project
source encoder & scrambler PSD (dB)
4000 Hz
May 5, 1999
LPF
PSD (dB)
-10
-20
-30
-40 0
4000 Hz
r(t) 2u(f )
bband
y(t) p(t)
y lpf
zk
z
ak
dec
reconstructed image symbol demapper descrambler & source bits decoder Xr
hilbert
e j2f0t
kT + t0
filter matched to training tone
corr
peak detector
t0
ix1
Warning: the automatic-gain control, carrier synchronization, and equalization systems are not shown!
