Enhanced receiver design for ultra-reliable low-latency short data communications

This thesis focuses on exploring important aspects of short data transmission, particularly the design of reliable and low-complexity receivers. Specifically, we are initially interested in designing advanced BICM receivers for short packets compliant with polar and LDPC coded transmissions. Indeed, the relevance of BICM becomes particularly evident in scenarios characterized by error-prone communication channels, requiring increased reliability. Its effectiveness also depends on underlying detection and decoding metrics, highlighting the importance of a nuanced trade-off between performance and complexity. The proposed receiver metrics consider scenarios where channel state information (CSI) is effectively unknown, aiming to assess the impact of various channel conditions. Hence, we present enhanced receivers for short data in the range of 20-100 bits for the envisaged beyond 5G/6G signaling scenarios by evaluating their performance over 5G short block channels, utilizing polar and LDPC coded formats. We look into receiver metrics exploiting joint estimation and detection}(JED), which is amenable to situations where low-density of DMRSs are interleaved with coded data symbols. We specifically address situations where accurate channel estimation is impossible, demonstrating that a well-conceived metric exploiting interleaved DMRS in the detection metric computation achieves performance comparable to a receiver with perfect channel state information. Remarkably, this approach demonstrates substantial performance gains when compared to conventional 5G OFDM receivers, applicable to both uplink and downlink transmission scenarios. In addition to this initial contribution, we also explore the aspect of designing reliable and low-complexity receivers compliant with Reed-Muller (RM) coded transmissions. The design of enhanced receivers capable of reliably detecting and decoding short packets in the range of 3 to 11 bits with lower complexity seems more than necessary to fully leverage their potential in URLLC and their use in NR signaling scenarios. Accordingly, we are focusing on the baseline 3GPP PUCCH in DMRS assisted transmission, specifically the use of Reed-Muller codes decodable via maximum-likelihood. Through the receiver structure, namely the estimator-correlator, we show that the non-coherent energy term, which is typically not used in conventional receivers, can result in a penalty that can significantly affect receiver performance in the operating regimes of 5G/6G systems. Furthermore, in current communication systems, the detection/decoding of short packets generally consists of channel estimation by the least squares method followed by quasi-coherent detection. However, it should be noted that in both cases, the underlying algorithms are suboptimal due to the detection procedure involving separate channel estimation followed by quasi-coherent detection. Conversely, the use of a fully non-coherent receiver would consequently incur a substantial complexity cost, especially when dealing with longer bit-length transmissions in the considered short block regime. Therefore, it is becoming realistic to consider alternative detection/decoding strategies that offer a favorable performance/complexity trade-off, compliant with such RM coded transmissions. Building on this challenge, we introduce the principle of block/segment encoding using First-order RM (FoRM) Codes, which are amenable to low-cost decoding through block-based fast Hadamard transforms (BFHT). The Block-based FHT has demonstrated to be cost-efficient with regard to decoding time, as it evolves from quadric to quasi-linear complexity with a manageable penalty in performance. Subsequently, incorporating an adaptive DMRS/data power adjustment, it becomes feasible to narrow the performance gap with respect to the conventional maximum likelihood receiver, leading to a good trade-off between performance and complexity to efficiently handle small payloads.