With the stringent requirements introduced by the new sixth-generation (6G) internet-of-things (IoT) use cases, traditional approaches to multiple access control have started to show their limitations. A new wave of grant-free (GF) approaches have been therefore proposed as a viable alternative. However, a definitive solution is still to be accomplished. In our work, we propose a new semi-GF coordinated random access (RA) protocol, denoted as partial-information multiple access (PIMA), to reduce packet loss and latency, particularly in the presence of sporadic activations. We consider a machine-type communications (MTC) scenario, wherein devices need to transmit data packets in the uplink to a base station (BS). When using PIMA, the BS can acquire partial information on the instantaneous traffic conditions and, using compute-over-the-air techniques, estimate the number of devices with packets waiting for transmission in their queue. Based on this knowledge, the BS assigns to each device a single slot for transmission. However, since each slot may still be assigned to multiple users, collisions may occur. Both the total number of allocated slots and the user assignments are optimized, based on the estimated number of active users, to reduce collisions and improve the efficiency of the multiple access scheme. To prove the validity of our solution, we compare PIMA to time-division multiple-access (TDMA) and slotted ALOHA (SALOHA) schemes, the ideal solutions for orthogonal multiple access (OMA) in the time domain in the case of low and high traffic conditions, respectively. We show that PIMA is able not only to adapt to different traffic conditions and to provide fewer packet drops regardless of the intensity of packet generations, but also able to merge the advantages of both TDMA and SALOHA schemes, thus providing performance improvements in terms of packet loss probability and latency.

翻译：随着第六代(6G)物联网(IoT)应用场景提出的严苛要求，传统的多址接入控制方法开始显现其局限性。因此，新一代免授权(GF)方法被提出作为可行的替代方案。然而，最终解决方案仍有待完善。在我们的工作中，提出了一种新的半免授权协调随机接入(RA)协议，称为部分信息多址接入(PIMA)，以降低数据包丢失和时延，特别是在偶发激活的场景下。我们考虑机器类通信(MTC)场景，其中设备需要在上行链路向基站(BS)传输数据包。使用PIMA时，BS能获取瞬时流量状况的部分信息，并利用空中计算技术估计队列中有数据包等待传输的设备数量。基于这一认知，BS为每个设备分配一个时隙进行传输。但由于每个时隙仍可能分配给多个用户，碰撞可能发生。基于估计的活跃用户数，对分配的时隙总数和用户分配进行联合优化，以减少碰撞并提升多址接入方案的效率。为验证方案有效性，我们将PIMA与时分多址(TDMA)和时隙ALOHA(SALOHA)方案进行对比——这两种分别是低流量和高流量条件下时域正交多址接入(OMA)的理想方案。结果表明，PIMA不仅能适应不同流量条件，无论数据包生成强度如何都能提供更少的丢包率，还能融合TDMA和SALOHA方案的优势，从而在丢包概率和时延方面带来性能提升。