Short-Length Channel Codes Design for Machine-to-Machine Communication

Machine-to-Machine (M2M) communications are expected to play an essential role in the upcoming fifth generation (5G) cellular networks. It is going to support various emerging services with new types of requirements including support of massive numbers of connected devices, high link reliability, and low latency real-time operations. To fulfill these requirements, the frame length of each individual transmission tends to be very small. This gives rise to new challenges of designing channel codes for short information length. On the basis of nested convolutional codes, we propose a new method of constructing rate-compatible tail-biting convolutional codes (RC-TBCC). This method allows us to search for a class of rate 1/n TBCC codes, which have a tradeoff between implementation flexibility and error performance for different information length. The numerical results show that the RC-TBCC codes constructed by our proposed method not only have near optimal error performance for long information length, but also have consistent error performance for short information length. Furthermore, our proposed codes outperform the TBCC codes used in the LTE standards. We also investigate the error performance of polar codes for short information length. Polar codes are designed for approaching channel capacity for long information length. Hence, they have performance degradation for short information length. Our simulation results show that the error performance of polar codes for short information length is not as good as TBCC codes. But with the aid of CRC, the error performance of polar codes for short information length can be greatly improved