Improved Frameless ALOHA for Wireless Networks

In uncoordinated multiuser transmission, each user transmits data multiple times uniformly at random, denoted as a repetition rate of the users. In particular, the transmission by users in the existing random access frameless ALOHA scheme follows the binomial distribution. As a result, the users have a large range of repetition rates. The repetition rate of a user is analogous to the degree of a variable node in the corresponding bipartite graph. Thus, the conventional frameless ALOHA scheme suffers from high error floor due to the lower repetition rates of users. Similarly, high transmit powers at the users are attributed to the higher repetition rates. To circumvent these problems, we propose a design framework based on the tools of graph-based codes to optimize the transmission strategy of users. The resulting transmission strategy eliminates the users with very low and very high repetition rates. Consequently, the proposed frameless ALOHA scheme demonstrates significant performance improvement especially in the error floor region. Furthermore, the users in the designed frameless ALOHA scheme require low transmit power as compared to its conventional counterpart. Numerical examples are provided to confirm the improved performance of the proposed scheme.