Page 99 - ITU Journal Future and evolving technologies Volume 2 (2021), Issue 1
P. 99
ITU Journal on Future and Evolving Technologies, Volume 2 (2021), Issue 1
(a)
10 0
(a)
Model 2: Parallel Hamming 2xCR4
-1
10
10 0
Model 3: Parallel Hamming 5xCR 2
10 -2 -1 Fit Model 3
10
BER 10 -3 10 -2
-4
BER 10 -3
10
10 -5 -4
10
-6
10 -5
0 2 4 6 8 10 12 10
E /N (dB)
b 0
-6
(b) 10
0 2 4 6 8 10 12
E /N (dB)
b 0
Fig. 8 – Model 2: (a)Short‑frame OFDM communication model using
MATLAB‑Simulink tools (b) BER performance plotted on BERTool ap‑ (b)
plication
Fig. 9 – Model 3: (a)Short‑frame OFDM communication model using
Model 2: Parallel Hamming code 2*[31,26] MATLAB‑Simulink tools (b) BER performance plotted on BERTool ap‑
plication
In order to improve the correction capacity of Model 1: 4.2 Comparison between parallel Hamming
Simple Hamming code Model [63, 57]. Here, we propose and simple Reed‑Solomon
to implement 2 couples of encoders/decoders in parallel
respectively, so we can improve the correction capability In this subsection, we will compare the previous mod‑
of the Hamming code; which will be 2 bits out of 52 bits els that are adapted to our technical requirements: Ham‑
(20% of the total message), and detection would be 4 out ming (Model 1, Model 2, Model 3) to Reed‑Solomon
6
of 52, while the irst model could only detect 2 and correct (Model 4). We want to verify if there is a signi icant differ‑
1. We use MATLAB‑Simulink tools in order to model the ence in BER performance: since parallel Hamming coding
parallel Hamming code communication system 2×[31, 26] is very interesting in terms of simplicity and robustness,
as shown in Fig. 8. however Reed‑Solomon is very interesting in terms of er‑
ror correction capability.
Model 3: Parallel Hamming code 5*[15,11] In this scenario, we choose Model 4 where = 56/64
6
is the simple Reed‑Solomon coding rate (around 50 bits).
For Model 3, we can cut the message on 5 times, each of Then we repeat the same simulations done for Model 1,
= 11 bits, which would be coded by 5 encoders Ham‑ Model 2 and Model 3 and for Reed‑Solomon (Model 4)
ming 5 × [15, 11]. Thus, the capacity of correction in this several times for each model and we calculate the aver‑
case would be of 5 bits out of 55 (10% of the total mes‑ age points for each value of 0 . We plot then the results
sage), and detection would be 10 out of 55, while the irst in Fig. 10.
model could only detect 2 and correct 1. We use MATLAB‑ As we can see in Fig. 10, the parallel coding represents a
Simulink tools in order to model the parallel Hamming gain in correction capacity, and both Model 2 and Model 3
code communication system 5 × [15, 11] as shown in Fig. converge faster to no errors than its equivalent with sim‑
9. ple Hamming (Model 1). In our application case, and es‑
© International Telecommunication Union, 2021 83