2.16  Exercises


2.1. An FSK modem uses eight distinct frequencies for signaling. How many symbols per second (in bauds) this modem must transmit to reach a 2.4 kbps bit rate?
2.2. Assume a digital binary signal m(t) in which the duration of 1 bit is equal to 1 second, modulating a sinusoidal carrier c(t) of frequency 6 Hz and peak amplitude of 1 V. For m(t) representing a sequence of five bits {1 1 0 0 1} using rectangular pulses, draw carefully the output waveforms m(t)c(t) of two modulators: (a) ASK with m(t) having amplitudes 1 and 0, (b) PSK with m(t) having amplitudes 1 and 1.
2.3. Being tired of symmetric constellations, a student designed {5,1,3,4} as its PAM constellation. Considering that the shaping pulse p(t) has unitary energy and the symbols, transmitted at 4 bauds, are independent and uniformly distributed: a) what is the average energy of a transmit symbol? b) what is the transmit signal power? c) can you criticize the chosen constellation? If yes, in what aspects?
2.4. A binary communication system uses antipodal signals p(t) and p(t), with p(t) = Asin (πtTsym) with 0 t Tsym or 0 otherwise, and A {±1}. Find a reasonable value for Tsym assuming the symbol rate must be 400 bauds. For the chosen value of T, sketch the transmit waveform corresponding to the information bits [0 1 0 0 1 0], where bit 1 corresponds to A = 1 and bit 0 to A = 1.
2.5. The generation of a 4-PAM discrete time signal s[n] is modeled by the convolution between the symbols m[n ] upsampled by a factor of L and a shaping pulse p[n]. Assume p[n] = 3δ[n] + 4δ[n 1] + 5δ[n 2], L=4, and symbols {3,1,1,3} with Gray mapping to bits {00,01,11,10}, respectively. Draw carefully the signal s[n] for the transmit bitstream 11 01 10 11 00.
2.6. Table 2.1 and the associated Figure 2.8 and Figure 2.7 contrast Gray and natural coding for 8-PAM. Assume the noise at the receiver has a relatively low power and that in all errors, a symbol is confused by one of its neighbor symbols. The symbols are equiprobable and the probability of symbol error is Pe = 0.01. Estimate the bit error probability Pb for each constellation.
2.7. Assume the 8-PAM constellation of Figure 2.8 is in volts and the symbols 7 and 7 have probability 0.3 each, while the remaining symbols are equiprobable. The basis function p(t) has unitary energy and the transmitted signal s(t) is obtained by multiplying the constellation symbols by p(t). The symbol period is Tsym = 2 s. Calculate: (a) the average power of s(t) and (b) the number of bits per second.
2.8. What are the theoretical power values that should be obtained with Listing 2.3 (consider an infinite number N of PAM symbols) for the sequences: a) m, b) m_upsampled and c) s.
2.9. An M-PAM system uses Rsym = 500 bauds and a shaping pulse p(t) with energy Ep such that the transmit PAM signal power P coincides with the average energy constellation E¯c. Design this constellation to maximize the bit rate constrained to use a transmit power of 25 dBm and a modulation order M that is a power of 2. The symbols must be uniformly spaced by a distance d. The path loss is 15 dB and the noise power at the receiver is such that the separation between neighboring PAM symbols at the receiver must be at least Δ = 60 mV to achieve the desired BER of 103. Inform: a) the number M of symbols, b) the respective bit rate, c) the corresponding E¯c and d) the value of each symbol in volts. You may find useful that the M-PAM constellation ± d 2,±3d 2 ,±5d 2 ,,±(M1)d 2 has an average energy E¯c given by Eq. (2.10).
2.10. Sketch the block diagram of a binary FSK digital communication system using only analog signal processing (do not use DSP, ADC nor DAC). Provide a high-level description using a block diagram with e. g. filters and amplifiers.
2.11. Using unsigned bytes (i. e., from 0 to 255): a) inform the bitstream corresponding to unpacking a 16-PAM organized as the bytes sequence 0, 255, 15, 1.
2.12. Create a constellation for 16-PAM with zero-mean and average energy of Ec = 4.5 J. You can list the code or the obtained symbols.
2.13. Modify the code of Application 2.5 to use b = 8 bits per symbol and include an AWGN channel in your simulation. Pick a file for testing, with size larger than 5 MB. Assuming Rsym = 100 bauds, what is the minimum SNR that you still achieve error-free transmission? (if you do not have diff, transmit a zip file and check if you can uncompress the received version).
2.14. Assume that a binary FSK system must be implemented with synchronous modulation using digital signal processing. The receiver has the following specifications: two frequencies in the uplink are 980 and 1180 Hz, the sampling rate is Fs = 9.6 kHz and Rsym = 3000 bauds, such that three samples represent each basis function. Assume the code below is used to obtain the two corresponding basis functions and explain: a) if they are orthogonal or not, b) the Octave/Matlab commands to normalize them for having unitary energy and check if they are orthonormal.

1T_sample = 1/Fs; t = 0:T_sample:1/Rsym-T_sample; %discretized time 
2basis0=cos(2*pi*980*t); basis1=cos(2*pi*1180*t); %basis functions

2.15. Assume a NRZ polar code with a pulse of amplitude A = 5 V, symbols with values B = ±1 V and Tsym = 1 ms, such that the PSD is Ss(f) = A2B2Tsymsinc2(fTsym). Provide detailed graphs of the shaping pulse’s Fourier transform magnitude |P(f)| and Ss(f).
2.16. Obtain an analytical expression for the PSD of the Manchester code used in the first generation of IEEE 802.3 Ethernet standards such as 10BASE-T. Assume ± 2.5 V as the voltage levels.
2.17. Generate a version of Figure A.26 describing the autocorrelation of an upsampled cyclostationary polar signal mu[n] with L = 3.
2.18. A coworker designed a new line code and your task is to calculate its PSD. The code has M = 4 equiprobable symbols: 3,1,1 and 3. The shaping pulse p(t) returns to zero (RZ): it has amplitude A over the interval [0,Tsym2] and zero amplitude over ]Tsym2,Tsym[, where Tsym is the symbol period. Provide an analytical expression to the corresponding PSD.