The 5G channel code contenders
Robert G. Maunder · ePrints Soton (University of Southampton) · 2016
Channel coding is a vital but complex component of cellular communication systems, which is used for correcting the communication errors that are caused by noise, interference and poor signal strength.The turbo code was selected to provide channel coding in 3G and 4G cellular systems, but the 3GPP standardization group is currently debating whether it should be replaced by the Low Density Parity Check (LDPC) or polar code in 5G.In this white paper, we dispel two myths that have been the genesis of this debate, namely that turbo codes cannot support high throughputs and that they have higher complexities than LDPC and polar codes.Indeed, AccelerComm has already demonstrated that turbo codes can achieve decoded throughputs exceeding the 5G target of 20 Gbps.Furthermore, this white paper shows that the overall implementation complexity of turbo, LDPC and polar codes depends on more than just their computation complexity.Since turbo codes have the benefits of low interconnect complexity and inherent flexibility, we show that they facilitate high-throughput flexible channel coding at lower implementation complexities than LDPC and polar codes.Furthermore, turbo codes offer the additional benefit of backwards compatibility to 3G and 4G, which offers the cellular communications industry some very significant cost savings.It is for these reasons that AccelerComm is promoting the turbo code for 5G. I. WHAT IS CHANNEL CODING?In cellular communication systems, wireless transmission is used to convey data between handsets and basestations, where the latter act as gatekeepers to the Internet and telephone networks.However, the received data typically differs to the transmitted data, owing to communication errors caused by noise, interference or poor signal strength.In order to correct these communication errors, cellular communication systems use channel codes.More specifically, a channel encoder is used in the transmitter (be it the handset or the basestation) to convert each so-called data block comprising K data bits into a longer encoded block comprising N > K encoded bits, which are transmitted.In the receiver, the additional (N -K) encoded bits provide the channel decoder with information that allows it to detect and correct communication errors within the original K data bits.If the noise, interference or poor signal strength is particularly severe, then a low coding rate R = K/N will be required for the channel decoder to successfully detect and correct all transmission errors.However, a low coding rate implies the transmission of a high number N of encoded bits, which consume transmission energy and bandwidth.Therefore, good channel codes are ones which allow the successful detection and correction of transmission errors at coding rates R that are as close as possible to the theoretical limit that is imposed by the channel capacity [1].In the past couple of decades, several nearcapacity channel codes have emerged, including the turbo codes that are adopted in 3G and 4G mobile broadband standards, the Low Density Parity Check (LDPC) codes [2] that are adopted in WiFi standards [3], as well as the more recent polar codes [4].Turbo and LDPC codes employ an iterative decoding process, in which each successive decoding attempt informs the next, until the process converges.By contrast, polar codes use a successive cancellation decoding process, in which the decoding of each successive bit informs the decoding of the next.Since the channel decoder must overcome the uncertainty introduced by noise, interference and poor signal strength, it typically has a much greater complexity than the channel encoder.Owing to this, it is the channel decoder that is typically the main concern when designing a channel code.