Journal on Communications ›› 2021, Vol. 42 ›› Issue (9): 31-42.doi: 10.11959/j.issn.1000-436x.2021155

• Papers • Previous Articles     Next Articles

Low error floor LT coding algorithm by using fixed-length node classification window

Xin SONG1, Naiping CHENG2, Shuyan NI2, Yurong LIAO2, Tuofeng LEI1   

  1. 1 Department of Graduate Management, Space Engineering University, Beijing 101416, China
    2 Department of Electronic and Optical Engineering, Space Engineering University, Beijing 101416, China
  • Revised:2021-06-30 Online:2021-09-25 Published:2021-09-01
  • Supported by:
    The National Natural Science Foundation of China(61805283)

Abstract:

To solve the problem of high error floor for LT codes in additive white Gaussian noise (AWGN) channels, an improved coding scheme was designed.A fixed-length classification window was used to sort out the information nodes with relatively small degrees.Then these nodes would be forced to participate in the coding process frequently, so that they could obtain sufficiently high reliability.To design the optimal parameters, firstly, the bit error rate (BER) performance of LT codes and the cause of the error floor were analyzed, and the theoretical BER lower bound was introduced as the first constraint.Secondly, the extrinsic information gain loss ratio (GLR) was designed to measure the convergence behavior of the improved LT code, and maximizing the GLR was denoted as the second constraint.Finally, the influence of different parameters on algorithm efficiency and coding complexity were analyzed, and the principle of priority selection of parameters was pointed out as the third constraint.The simulation results show that compared with the conventional LT code, the proposed algorithm can reduce the error floor by nearly three orders of magnitude.Besides, the designed LT code outperforms the existing LT codes in literature in terms of bit error rate (BER) performance.

Key words: channel coding, fountain code, convergence, bit error rate

CLC Number: 

No Suggested Reading articles found!