This is the homework from COM5170 Wireless Communication in National Tsing Hua University. We are going to implement several diversity combining strategies. The requirements are shown as following.
There are L (L = 1, 2, 3, 4) diversity branches of uncorrelated Rayleigh/Ricean fading signals. The average symbol energy-to-noise power ratio Es/N0 of each branch is 1, 3, 5, 7, and 9 dB. Simulate the QPSK bit error rate for
- Selective Combining
- Maximal Ratio Combining
- Equal Gain Combining
- Direct Combining
Select the branch with highest signal-to-noise from received signal.
SNR of received signal:
$$ SNR=\frac{|g_k |^2 E_s}{E[n^2 ]}, k=1-L $$
Because
Next, compensate the phase shift
The diversity branches are weighted by their complex fading gains and then combined. $$ r(t)=\sum _{k=1} ^L g_k ^* \tilde{r}(t)=\sum _{k=1} ^L g_k ^* g_k \tilde{s}(t)+\sum _{k=1} ^L g_k ^* \tilde{n}(t)=\sum _{k=1} ^L |α_k |^2 \tilde{s}(t)+noise $$
Since QPSK has equal energy symbols, EGC is useful. The diversity branches are not weighted. We compensate the phase shift and combine all branches. $$ r(t)=\sum _{k=1} ^L e^{-j\phi _k } \tilde{r} _k (t)=\sum _{k=1} ^L e^{-j\phi _k } g_k \tilde{s}(t)+\sum _{k=1}^L e^{-j\phi _k } \tilde{n}_k (t)=\sum _{k=1} ^L \alpha _k \tilde{s}(t)+noise $$
Combine all signals of branches directly and then compensates the overall phase shift. $$ r(t)=e^{-j\phi } \sum _{k=1} ^L \tilde{r} _k (t), φ=∡(\sum _{k=1} ^L \tilde{r} _k(t)) $$
The following 2 figures are both use 4 branches.
- Performance: MRC > EGC > SC > DC
The following figure use Equal gain combining with 4 branches.
- Performance: Ricean > Rayleigh