Joint Optimization of OFDM and Channel Coding for URLLC in Industrial Channels
Yasantha Samarawickrama,lvaro A. M. de Medeiros,Victor Cionca,Alvaro A. M. de Medeiros
DOI: https://doi.org/10.1109/tii.2022.3214513
IF: 12.3
2022-01-01
IEEE Transactions on Industrial Informatics
Abstract:Industrial communications have very tight requirements of reliability and latency, which are challenging to achieve at highly dynamic industrial radio channels. Redundancy in the modulation and coding can improve reliability but incurs excess latency. Minimizing latency subject to reliability constraint for a given channel state opens up the scope for static and runtime optimization. In this article, we explore joint optimal configuration of orthogonal frequency division multiplexing (OFDM) and channel coding, using polar, low density parity check (LDPC) and convolutional codes to minimize the transmission time subject to a maximum acceptable packet error rate, for three realistic channels reported by National Institute of Standards and Technology, using channel impulse response measurements, as well as perfect and imperfect channel estimation. The results show that the use of strong channel codes can significantly reduce transmission time, by up to 24 compared to convolutional codes, with LDPC achieving fastest transmission. For transmission times of 1020 ${\rm\mu}$s, the contribution of the cyclic prefix (CP) is significant, up to 30 for convolutional codes. Stronger channel codes allow reduced CP, even considering the added redundancy such as cyclic redundancy check, resulting in overall lower transmission time. The decoding time also introduces significant overheads, up to 50 of transmission time for polar codes. The article also investigates the impact of channel estimation errors, which adversely impact ultrareliable low-latency performance: Lower bound errors violate the reliability constraint and higher bound errors introduce excess transmission times.
automation & control systems,computer science, interdisciplinary applications,engineering, industrial