The transition to AI-native 6G networks embeds artificial intelligence (AI) across all network functions, enabling automation, efficiency, and advanced services. The AI-native approach is more than just a technical enhancement; it is a fundamental shift that drives the evolution toward autonomous networks, unlocking new advancements in network capabilities and user experience. A key enabler is AI as a Service (AIaaS), which allows Communication Service Providers (CSPs) to expose AI capabilities via standardized APIs. Built on Machine Learning Operations (MLOps), AIaaS streamlines AI model lifecycle management, reducing complexity for developers and optimizing AI-driven workflows.This paper investigates the integration of AIaaS into network exposure frameworks such as the 3rd Generation Partnership Project (3GPP) Network Exposure Function (NEF), Service Enabler Architecture Layer (SEAL), and the open-source CAMARA interface emphasizing the interacting and collaborating robots (cobots) use case in the context of industrial automation. We discuss how AIaaS enables CSPs to offer AI-powered services with guaranteed Quality of Service (QoS), driving innovation across industries and shaping the future of AI-native 6G networks.

Internet of Mobile Things (IoMTs) refers to the interconnection of mobile devices, for example, mobile phones, vehicles, robots, etc. For mobile data, strong extra processing resources are normally required due to the limited physical resources of the mobile devices in IoMTs. Due to latency or bandwidth limitations, it may be infeasible to transfer a large amounts of mobile data to remote server for processing. Thus, distributed computing is one of the potential solutions to overcome these limitations. We consider the device mobility in IoMTs. Two situations of the movement position of the mobile devices, i.e., unpredictable and predictable, are considered. In addition, three possible relative positions between the two server sets which respectively correspond to the positions of a mobile device for computation tasks offloading and for output results receiving, i.e., within the same server sets, with two different server sets and with two adjacent server sets, are studied. Coded schemes with high flexibility and low complexity are proposed based on Fountain codes to reduce the total processing time and latency of the distributed fog computing process in IoMTs for the above different situations. The latency related performance, i.e., the computation, the communication and the transmission loads, is analyzed. We also compare of the Fountain code-based and the uncoded schemes and numerical results demonstrate that shorter total processing time and lower latency can be achieved by the Fountain code-based schemes.

In this paper, we study the characteristics of dominant interference power with directional reception in a random network modelled by a Poisson Point Process. Additionally, the Laplace functional of cumulative interference excluding the n dominant interferers is also derived, which turns out to be a generalization of omni-directional reception and complete accumulative interference. As an application of these results, we study the impact of directional receivers in random networks in terms of outage probability and error probability with queue length constraint.

In Industrial Internet of Things (IIoTs), data generated during manufacturing are collected by sensors and need be processed timely. The direction of data transmissions from sensors to processing centers (fog nodes) is often called uplink transmission. In this paper, the cases with single and multiple distributed fog nodes, which are also referred to as centralized and distributed models, are studied. Two distributed schemes based on batched sparse (BATS) codes are proposed separately for the uplink of these two models. The expected rank and the recovery probability of the information from sensors at fog node(s) are derived. Comparison results show that by using the proposed BATS-based schemes, improved transmission reliability can be achieved compared to the XOR-based network coding (NC) scheme.

Content caching at small base stations has been considered as an efficient way to alleviate the use of expensive backhauling. In this paper, we study the caching content update for a cache- enabled heterogeneous network. Through analyzing a benchmark content update policy, the problem of content consistency, which is caused by the distinct update time at each caching entity, is firstly revealed. As well, the close form representation of the consistency probability in a mobile environment is derived. Furthermore, a deterministic content update strategy is investigated, where the trade-off between consistency probability and storage cost is characterized. Detailed simulations are provided to support our analysis, as well as present the impacts of moving speed of mobile terminals and transmission bandwidth for content delivery on the consistency performance.

A social network is a social structure made up of a set of social actors and a set of dyadic ties between these actors. The actors form a number of communities. In communities, some actors want to transmit their information to all other actors. Each actor corresponds to a user equipment (UE). The UE of the actor which has information to be transmitted is also called the source, and the UEs of all other actors are called destinations. The information is transmitted from the source to destinations with the assistance of helpers, which can be small cell base stations (SCBSs). A novel information dissemination approach, namely expanded network coding with a finite buffer size (ENCFB), is proposed for the case when the buffer size of helpers is limited. The performance comparison of the uncoded information dissemination approach, the network coded approach and the ENCFB approach is conducted. Comparison results show that the ENCFB approach can significantly improve the performance of information dissemination when the buffer size is limited.

In an industrial automation system, one of the most important parts is control loop. Fog computing is a potential solution for industrial control in time-critical applications as it provides distributed computing services closer to the connected devices. However, a huge amount of data exchanging among fog nodes causes high communication load, which constrains the overall response time from fog nodes to actuators. In this paper, we consider the erasure environment, batched sparse (BATS) codes are applied to the Map and the Data Shuffling stages of distributed fog computing process to reduce both the communication and the computation loads. The communication loads of the uncoded, the coded, and the proposed BATS-based schemes over erasure channels are calculated, respectively. Numerical results show that the BATS-based scheme can reduce the communication and the computation loads simultaneously, and furthermore reduce the overall response time from fog nodes to the actuators.

In this paper, we investigate the performance of a wireless sensor network, in which multiple groups of source nodes communicate with their respective destination nodes with the help of a common relay network. A network code division multiplexing (NCDM) scheme is proposed to remove the inter-session interference among multiple transmission sessions at each destination. We focus on analyzing the soft processing algorithm of the NCDM scheme. Based on the analysis results, a new code design criteria for the construction of the generator matrix is proposed. Simulation results show that by following the proposed code design criteria, the bit error ratio (BER) performance gap between the scheme we studied and the serial session scheme can be managed effectively. In serial session scheme, source nodes in a number of groups communicate with their respective destinations in a time division manner.

In this paper, we investigate the performance of a wireless relay network with multiple transmission sessions, in which multiple groups of source nodes communicate with their respective destination nodes via a shared wireless relay network. A multiple transmission session model with network code division multiplexing (NCDM) scheme is proposed to remove the inter-session interference at each destination. The fundamental idea of the NCDM scheme takes advantage of the property of G Θ HT= 0 of the low-density generator matrix (LDGM) codes. Based on the analysis of the NCDM scheme, we investigate the relationship among the equivalent received signal vector, the number of sessions and the column weight of the generator matrix. New code design criteria for the construction of the generator matrix is proposed. We further evaluate the multiple transmission session model with the proposed NCDM scheme in terms of throughput and complexity. Our evaluation demonstrates that the proposed scheme not only has a linear computational complexity, but also shows a similar error performance in the AWGN case and a considerable throughput improvement compared with its counterpart, which is referred to as a serial session scheme, where groups of source nodes communicate with their respective destinations in a time division manner.

In this paper, we first analyse bit error rate (BER) bounds of the distributed network coding (DNC) scheme based on the Luby-transform (LT) codes, which is a class of fountain codes, for wireless sensor networks (WSNs). Then we investigate the effect from two parameters of the degree distributions, i.e., the degree value and the proportion of odd degree, to the performance of the LT-based DNC scheme. Based on the analysis and investigation results, a degree distribution design criteria is proposed for the DNC scheme based on fountain codes over Rayleigh fading channels. We compare the performance of the DNC scheme based on fountain codes using degree distributions designed in this paper with other schemes given in the literature. The comparison results show that the degree distributions designed by using the proposed criteria have better performance.

In this paper, we propose a distributed network coding (DNC) scheme based on the Raptor codes for wireless relay networks (WRNs), where a group of source nodes communicate with a single sink through a common relay network in a multi-hop fashion. At the sink, a graph-based Raptor code is formed on the fly. After receiving a sufficient number of encoded packets, the sink begins to decode. The main contributions of this paper are the derivations of upper and lower bit error rate (BER) bounds for the proposed Raptor-based DNC scheme.

In many applications of wireless relay networks (WRNs), such as image and video systems, unequal error protection (UEP) among the transmission data from different source nodes is required. All the source nodes in a WRN with different protection requirements form multiple protection groups. In this paper, we focus on estimating the protection parameter, i.e., the protection weight, for each protection group. We first analyze the bit error rate (BER) upper bound of the UEP distributed fountain codes over Rayleigh fading channels for WRNs. Then based on the analysis results, we derive the approximate expression for the BER upper bound with protection weight as the variable. The derived approximate expression can be used to estimate the protection weights for the protection groups according to their various performance requirements. Finally, examples on the application of the approximate expression in estimating protection weights are given.

In wireless relay networks (WRNs), multiple source nodes communicate with a single destination node through a common relay network. Different source nodes may have different error protection or recovery time requirements. In this paper, we focus on unequal error protection (UEP) distributed network coding (DNC) design for WRNs. Specifically, we propose a continuous UEP method based on fountain codes with the objective of satisfying the UEP or unequal recovery time (URT) requirements. Then based upon this UEP method, we develop an adaptive UEP DNC scheme to realize adaptive UEP in WRNs. We analyze the properties of the proposed adaptive UEP DNC scheme and derive the upper and lower bit error rate (BER) bounds for it over Rayleigh fading channels under maximum-likelihood (ML) decoding. Simulation results show that the proposed adaptive UEP DNC scheme has desirable UEP and URT properties. The transmitted data from the source nodes in different protection groups can be recovered successfully under their performance requirements in a shorter time by using the adaptive UEP DNC scheme.

In this paper, we focus on network coding design for the wireless sensor networks (WSNs), where multiple source nodes communicate with a common destination node with the help of multiple relay nodes in a two-hop fashion. Specifically, we propose an unequal error protection (UEP) distributed network-channel coding (DNCC) scheme based on Luby-transform (LT) codes. We analyse three properties of the proposed UEP DNCC scheme, i.e. effective weights, turning points, and thresholds of the source nodes' number. Also, we derive the upper and lower bit error rate (BER) bounds for the proposed UEP DNCC scheme over Rayleigh fading channels under maximum-likelihood (ML) decoding. Based on the analysis, it is observed that the proposed UEP DNCC scheme can achieve all protection levels required when the number of source nodes is large enough. Simulation results show that our UEP DNCC scheme can provide desirable UEP to all source nodes.

In this paper, we propose a distributed network coding (DNC) scheme based on the Raptor codes for wireless sensor networks (WSNs), where a group of sensor nodes, acting as source nodes, communicate with a single sink through some other sensor nodes, serving as relay nodes, in a multi-hop fashion. At the sink, a graph-based Raptor code is formed on the fly. After receiving a sufficient number of encoded packets, the sink begins to decode. The main contributions of this paper are the derivation of a bit error rate (BER) lower bound for the LT-based DNC scheme over Rayleigh fading channels under maximum-likelihood (ML) decoding, and the derivations of upper and lower BER bounds for the proposed Raptor-based DNC scheme on the basis of the derived BER bound of LT codes.

In this paper, we analyse the performance of distributed network-channel coding (DNCC) with multiple destinations executing a code nulling (MDCN) process. By analysing the formulation deduced from DNCC with the MDCN process, we find that some unstable zero elements and additional noise are generated after right multiplying the parity-check matrix. These unstable zero elements and additional noise are the reasons to degrade BER performance in the two groups of source nodes and two destination nodes (TSTD) network model. Theoretical bit error ratio (BER) curves are drawn according to the calculated equivalent received signal-to-noise ratio (SNR). The analysis results are consistent with the theoretical curves. A design principle for the generator matrix of the DNCC scheme is proposed to solve the BER performance degradation problem. Simulation results show that the problem caused by the MDCN process can be managed effectively and the BER performance can be improved significantly by using the proposed design principle.