To support multiple AI-based applications, mobile systems need to collaboratively execute DNN architectures on heterogeneous AI accelerators. At the same time, the increasing DNN complexity and high degree of diversity in workloads on multichip module (MCM) accelerators are pushing AI processing off mobile nodes onto the edge. This has made computationally intensive, edge-based solutions the dominant approach for the deployment of modern neural networks. However, the rigid structure of fully-executed DNNs fails to align with the modular nature of MCM architectures, limiting their potential for efficient execution. In this paper, we introduce CLEAR, a novel optimization framework based on geometric programming that leverages both transformer-based and more canonical DNNs with early exits. CLEAR enables fast, coordinated decisionmaking across DNN design, workload distribution, and resource allocation, with the overarching goal of minimizing inference energy consumption. To our knowledge, this is the first work to integrate dynamic DNN optimization with decisions at both the communication infrastructure and hardware accelerator levels. We evaluate CLEAR using real-world wireless measurements and dynamic DNNs applied to computer vision inference tasks. Our results demonstrate that CLEAR achieves near-optimal performance and reduces energy consumption and resource usage by over 80% and 70%, respectively, compared to its benchmark.

Efficient network modeling is essential for resource optimization and network planning in next-generation large-scale complex networks. Traditional approaches, such as queuing theory - based modeling and packet-based simulators, can be inefficient due to the assumption made and the computational expense, respectively. To address these challenges, we propose an innovative energy-efficient dynamic orchestration of Graph Neural Networks (GNN) based model training and inference framework for context-aware network modeling and predictions. We have developed a low-complexity solution framework, QAG, that is a Quantum approximation optimization (QAO) algorithm for Adaptive orchestration of GNN-based network modeling. We leverage the tripartite graph model to represent a multi-application system with many compute nodes. Thereafter, we apply the constrained graph-cutting using QAO to find the feasible energy -efficient configurations of the GNN-based model and deploying them on the available compute nodes to meet the network modeling application requirements. The proposed QAG scheme closely matches the optimum and offers atleast a 50% energy saving while meeting the application requirements with 60% lower churn-rate.

Next-generation mobile systems will support multiple AI-based applications, each leveraging heterogeneous sensors and data sources through deep neural network (DNN) architectures collaboratively executed within the network. In this context, to minimize the cost of the AI inference task subject to requirements on latency, quality, and – crucially –reliabilityof the inference process, it is vital to optimize (i) the set of sensors/data sources and (ii) the DNN architecture, (iii) the network nodes executing sections of the DNN, and (iv) the resources to use. To achieve these goals, we leverage dynamic gated neural networks with branches, and propose a novel algorithmic strategy called Quantile-constrained Inference (QIC), based upon quantile-Constrained policy optimization. QIC makes joint, high-quality, swift decisions on all the above aspects of the system, with the aim to minimize inference energy cost. We remark that this is the first contribution connecting gated dynamic DNNs with infrastructure-level decision making. We evaluate QIC using a dynamic gated DNN with stems and branches for optimal sensor fusion and inference, trained on the RADIATE dataset offering Radar, LiDAR, and Camera data, and real-world wireless measurements. Our results confirm that QIC closely matches the optimum and outperforms existing approaches in reducing energy consumption (compute, communication, and total) and application requirements failure by over 70%.

Mobile edge computing (MEC) is a promising technology to meet the increasing demands and computing limitations of complex Internet of Things (IoT) devices. However, implementing MEC in urban environments can be challenging due to factors like high device density, complex infrastructure, and limited network coverage. Network congestion and connectivity issues can adversely affect user satisfaction. Hence, in this article, we use uncrewed aerial vehicle (UAV)-assisted collaborative MEC architecture to facilitate task offloading of IoT devices in urban environments. We utilize the combined capabilities of UAVs and ground edge servers (ESs) to maximize user satisfaction and thereby also maximize the service provider’s (SP) profit. We design IoT task-offloading as joint IoT-UAV-ES association and UAV-network topology optimization problem. Due to NP-hard nature, we break the problem into two subproblems: offload strategy optimization and UAV topology optimization. We develop a Three-sided Matching with Size and Cyclic preference (TMSC) based task offloading algorithm to find stable association between IoTs, UAVs, and ESs to achieve system objective. We also propose a K-means based iterative algorithm to decide the minimum number of UAVs and their positions to provide offloading services to maximum IoTs in the system. Finally, we demonstrate the efficacy of the proposed task offloading scheme over benchmark schemes through simulation-based evaluation. The proposed scheme outperforms by 19%, 12%, and 25% on average in terms of percentage of served IoTs, average user satisfaction, and SP profit, respectively, with 25% lesser UAVs, making it an effective solution to support IoT task requirements in urban environments using UAV-assisted MEC architecture.

An increasing number of mobile applications leverage deep neural networks (DNN) as an essential component to adapt to the operational context at hand and provide users with an enhanced experience. It is thus of paramount importance that network systems support the execution of DNN inference tasks in an efficient and sustainable way. Matching the diverse resources available at the mobile-edge-cloud network tiers with the applications requirements and the complexity of their, while minimizing energy consumption, is however challenging. A possible approach to the problem consists in exploiting the emerging concept of dynamic DNNs, characterized by multi-branched architectures with early exits enabling sample-based adaptation of the model depth. We leverage this concept and address the problem of deploying portions of DNNs with early exits across the mobile-edge-cloud system and allocating therein the necessary network, computing, and memory resources. We do so by developing a 3-stage graph-modeling method that allows us to represent the characteristics of the system and the applications as well as the possible options for splitting the DNN over the multi-tier network nodes. Our solution, called Feasible Inference Graph (FIN), can determine the DNN split, deployment, and resource allocation that minimizes the inference energy consumption while satisfying the nodes’ constraints and the requirements of multiple, co-existing applications. FIN closely matches the optimum and leads to over 89% energy savings with respect to state-of-the-art alternatives.

The increasing pervasiveness of intelligent mobile applications requires to exploit the full range of resources offered by the mobile-edge-cloud network for the execution of inference tasks. However, due to the heterogeneity of such multi-tiered networks, it is essential to make the applications’ demand amenable to the available resources while minimizing energy consumption. Modern dynamic deep neural networks (DNN) achieve this goal by designing multi-branched architectures where early exits enable sample-based adaptation of the model depth. In this paper, we tackle the problem of allocating sections of DNNs with early exits to the nodes of the mobile-edge-cloud system. By envisioning a 3-stage graph-modeling approach, we represent the possible options for splitting the DNN and deploying the DNN blocks on the multi-tiered network, embedding both the system constraints and the application requirements in a convenient and efficient way. Our framework – named Feasible Inference Graph (FIN) – can identify the solution that minimizes the overall inference energy consumption while enabling distributed inference over the multi-tiered network with the target quality and latency. Our results, obtained for DNNs with different levels of complexity, show that FIN matches the optimum and yields over 65% energy savings relative to a state-of-the-art technique for cost minimization.

Mobile systems will have to support multiple AI-based applications, each leveraging heterogeneous data sources through DNN architectures collaboratively executed within the network. To minimize the cost of the AI inference task subject to requirements on latency, quality, and - crucially - reliability of the inference process, it is vital to optimize (i) the set of sensors/data sources and (ii) the DNN architecture, (iii) the network nodes executing sections of the DNN, and (iv) the resources to use. To this end, we leverage dynamic gated neural networks with branches, and propose a novel algorithmic strategy called Quantile-constrained Inference (QIC), based upon quantile-Constrained policy optimization. QIC makes joint, high-quality, swift decisions on all the above aspects of the system, with the aim to minimize inference energy cost. We remark that this is the first contribution connecting gated dynamic DNNs with infrastructure-level decision making. We evaluate QIC using a dynamic gated DNN with stems and branches for optimal sensor fusion and inference, trained on the RADIATE dataset offering Radar, LiDAR, and Camera data, and real-world wireless measurements. Our results confirm that QIC matches the optimum and outperforms its alternatives by over 80%.

The Internet of Things (IoT) is a promising application for 5G networks because it provides connected vehicle networks with an optimal blend of cost, latency, and speed. The IoT devices involved with vehicles are called vehicular IoT (V-IoT), requiring a dependable wireless communication system for safe and efficient operation. Compared to conventional waveforms like orthogonal frequency division multiplexing, the orthogonal time frequency space (OTFS) is superior due to its delay-Doppler domain modulation in high-mobility circumstances of V-IoTs. Integrating the$N$-dimensional ($N$-D) mapper with the improved minimum Euclidean distance (MED) into the conventional OTFS modulation technique for V-IoT can considerably enhance the bit error rate (BER) at higher modulation orders. So, we present an ND-OTFS-based V-IoT system with the proposed system's detector complexity analysis and the minimum Euclidean distance of the$N$-D signal constellations. The simulation findings demonstrate that for higher modulation orders in various time-variant channels, the proposed ND-OTFS-based V-IoT system outperforms the conventional OTFS-based V-IoT system regarding BER performance.

Tele-operated driving enables industrial operators to control heavy machinery remotely. By doing so, they could work in improved and safe workplaces. However, some challenges need to be investigated while presenting visual information from on-site scenes for operators sitting at a distance in a remote site. This paper discusses the impact of video quality (spatial resolution), field of view, and latency on users' depth perception, experience, and performance in a lab-based tele-operated application. We performed user experience evaluation experiments to study these impacts. Overall, the user experience and comfort decrease while the users' performance error increases with an increase in the glass-to-glass latency. The user comfort reduces, and the user performance error increases with reduced video quality (spatial resolution).

Recent advancement in multimedia and communication network technology have made interactive multimedia and tele-operation applications possible. Teledriving, teleoperation, and video-based remote-controlling require real-time live-video streaming and are delay critical in principle. Supporting such applications over wireless networks for mobile users can pose fundamental challenges in maintaining video quality and service latency requirements. This paper investigates the factors affecting the end-to-end delay in a video-based, remotely controlled moving platform application. It involves the real-time acquisition of environmental information (visually) and the delay-sensitive video streaming to remote operators over wireless networks. This paper presents an innovative experimental testbed developed using a remote-controlled toy truck, off-the-shelf cameras, and wireless fidelity (Wi-Fi) network. It achieves ultra-low end-to-end latency and helped us in performing the delay, network, and video quality evaluations. Extending the experimental study, we also propose a real-time live-media streaming control (RTSC) algorithm that maximizes the video quality by selecting the best streaming (network, video, and camera) configuration while meeting the delay and network availability constraints. RTSC improves the live-streaming video quality by about 33% while meeting the ultra-low latency (< 200 milliseconds) requirement under constrained network availability conditions.

A 5G network’s use-case in the Internet of Things (IoT) is a breakthrough, offering networks the ability to handle billions of connected devices with the proper blend of speed, latency, and cost. The IoT networks implemented in high-speed scenarios like intra and intervehicular communications in autonomous driving vehicles, high-speed vehicles, and trains will experience a Doppler effect. Orthogonal frequency-division multiplexing, the popular transmission technology for existing new-radio IoT (NR-IoT), is limited in providing reliable connections in high-speed vehicular scenarios. The performance of such system degrades with higher-order antenna configuration due to the lack of channel state information in highly mobile environments. The recently proposed orthogonal time-frequency space (OTFS) modulation is a strong contender that can handle high mobility but requires efficient transceiver design to be deployed in vehicular NR-IoT (V-IoT) systems. To conserve the resources and minimize the air time of the devices, we have proposed an embedded pilot design in the Delay–Doppler domain for the V-IoT systems. In the designed frame structure, the pilot’s position is optimized as per the vehicle speed to maximize the spectral efficiency of the system. The increase in spectral efficiency is at the cost of interference in the channel search region of the received Delay–Doppler domain OTFS signal. So a new joint estimator and the low-complex detector are proposed to handle the interference. The proposed efficient transceiver design with the spectral efficient pilot patterns allows us to conserve resources and remove complex encoder–decoders like the low-density parity check in NR-IoT.

Wireless communication between an unmanned aerial vehicle (UAV) and the ground base station is susceptible to adversarial jamming. In such situations, it is important for the UAV to indicate a new channel to the BS. This paper describes a method of creating spatial codes that map the chosen channel to the location of the UAVs in space, wherein the latter physically traverses the space from a given so called ”constellation points” to another. These points create patterns in the sky, analogous to modulation constellations in classical wireless communications, and are detected at the BS through a millimeter-wave radar sensor. A constellation point represents a distinct n-bit field mapped to a specific channel, allowing simultaneous frequency switching at both ends without any RF transmissions. The main contributions of this paper are: (i) We conduct experimental studies to demonstrate how such constellations may be formed using COTS UAVs and mmWave sensors, (ii) We develop a theoretical framework that maps a desired constellation design to error and band switching time, including multi-user scenario-specific challenges, (iii) We compare our approach against current FHSS technology and (iv) We experimentally demonstrate jamming resilient communications and validate system goodput for links formed by UAV-mounted software defined radios.

Orthogonal Time Frequency Space (OTFS) modulation is a 2-D modulation technique where the time-varying multipath channel is equivalent to a time-invariant delay-Doppler channel. The information symbols that are coherently combined along the multiple delay-Doppler diversity branches experience the same channel gain. New Radio (NR) based Internet of Things (IoT) in 5G standard can support high-speed automated vehicular network that can be termed as vehicular IoT (V-IoT). At high speeds, the orthogonal frequency division multiplexing-based transmission suffers from high intercarrier (or symbol) interference, making it unsuitable. In this paper, we have proposed a less complex NR-based IoT (NR-IoT) system with OTFS that performs better than NR-IoT with low-density parity-check (LDPC). We thereby eliminate the computational complexity of LDPC decoding in our scheme. We have evaluated our proposed technique of uncoded NR-IoT with OTFS over the extended vehicular channel model-A and ultra-reliable low latency communication channel. We have achieved a 6 dB SNR gain with our proposed NR-IoT system with OTFS.

Multi Input Multi Output (MIMO) based Cognitive Radio (CR) is implementable using Software Defined Radio (SDR) and is promising for efficient use of scarce electromagnetic spectrum. The Federal Communications Commission has proposed to create Citizens Broadband Radio Service (CBRS) with the three-tier spectrum sharing system to release more spectrum for the mobile broadband usage in next-generation wireless network. In this paper we propose an efficient cooperative spectrum sensing solution, ECSS, for CBRS CR system. The Priority Access License (PAL) are represented as Primary Users (PU) and General Authorized Access (GAA) as secondary user (SU). Modelling and analysis of cooperative spectrum sensing (CSS) is an important aspect in CR systems in presence of multiple users. In our scheme, spectrum sensing is carried by individual SUs using energy detector and Square Law Selection diversity. This is due to the simplicity and non-necessity of apriori information in both these techniques. The energy detection (ED) of an unknown signal over Nakagami-m fading is considered for analytically modelling local probability of detection at individual SU, termed as efficient spectrum sensing (ESS). Individual SUs send the hard decision rule to the Spectrum Access System (SAS), which acts as the fusion centre of the cooperative spectrum sensing system, where majority combining decision rule is employed to take final sensing decision using a Efficient Cooperative Spectrum Sensing (ECSS) Algorithm. Closed-form solutions for the ECSS system-level probability of detection and false alarm at the SAS are obtained and have been shown to give efficient performance.

This paper presents an innovative multi-faceted architecture, called UE-TV, for user-centric, revenue aware, and energy-efficient TV broadcast. Scalable high-efficiency video coded high definition television (HDTV) content is broadcast over LTE multicast/broadcast single frequency network (MBSFN) along with the available Wi-Fi access points (APs). The proposed framework adaptively encodes the TV content and allocates radio resource based on current network. Stackelberg two-stage game theoretic approach discerns an optimal transmit power at the LTE base stations (eNodeBs) and the proportion of subscribers that are respectively served by eNodeBs and available Wi-Fi APs. Varied user equipment (UE) resolutions, users' energy/price sensitivities, and channel conditions govern the service options and user satisfaction. Our analysis and simulations show that, in comparison with the broadcast schemes over MBSFN without or with adaptive video coding, UE-TV framework significantly enhances the user satisfaction via optimized price/quality trade-off as well as energy saving at the eNodeBs and UEs.

With device-to-device (D2D) based relay-assisted communications, deep faded cellular users (CUs) get a better connectivity. The deep faded CU's achievable throughput of the link from base station (BS) to the CU is less than the rate demand from the CU (unsatisfied CUs). In this paper, we study the resource allocation problem for both CUs and D2D pairs to increase the achievable throughput. This is accomplished by reducing the difference between rate demand and rate achieved for the active CUs. The resource allocation optimization problem is solved in three steps. The initial set of resources that can be allocated to each active CUs is determined in the first step. In the second step, it finds the set of active CUs which needs D2D assistance and then it finds suitable D2D relay nodes (DRNs). Resource allocation to those DRNs and D2D pairs without degrading the link throughput of active CUs is performed in the third step. Numerical results shows significant amount of throughput gain for the active CUs using D2D communication. We also measure the performance of our proposed model in terms of churn rate and the number of excess resource blocks (RBs) in the system.

Efficient Multimedia Broadcast and Multicast Services (MBMS) to heterogeneous users in cellular networks imply adaptive video encoding, layered multimedia transmission, optimized transmission parameters, and dynamic broadcast area definition. This paper deals with MBMS by proposing a multi-dimensional approach for broadcast area definition, which provides an effective solution to all of the above aspects. By using multi-criteria K-means clustering, our scheme provides users with high levels of Quality-of-Experience (QoE) of multimedia services. Adaptive video encoding and allocation of radio resources (i.e., time-frequency resource blocks, and modulation and coding scheme) are performed based on user spatial distribution, channel conditions, service request, and user display capabilities. Simulation results show that our solution provides a 70% improvement in user QoE and 86% in number of served customers, as compared to an existing multimedia broadcast scheme.

This paper presents an innovative multi-faceted architecture, named U-TV, that provides user-centric and adaptive digital television (DTV) broadcast for heterogeneous users over heterogeneous wireless networks. The service providers (SPs) are the DTV base station (DTV BS) and Wi-Fi access points (Wi-Fi APs). The Bertrand duopoly game theoretic approach determines the pricing policy of the SPs and also the proportion of subscribers each of these SPs serve. The proposed framework incorporates the user-end as well as system utility definitions based on the users' energy-saving and price trade-off. The joint-optimization solution facilitates scalable video encoding subject to the device display resolutions, and energy and price sensitivities. The proposed solution results in an increased user satisfaction by providing them with optimized video content over the price competent SPs. Additionally, the SPs are able to serve more customers with acceptable user experience, as compared to the conventional as well as adaptive DTV service networks.

This paper presents an adaptive SVC (scalable video coding) multimedia broadcast framework for the heterogeneous wireless mobile receivers with varying display resolutions, battery constraints, channel conditions, and service subscription priorities. In a bid to achieve optimal performance two variants of optimizations, namely, Subscriber-count Maximizing Adaptive SVC Rate and Energy Saving (SM-ASRES) scheme and Revenue Maximizing ASRES (RM-ASRES) are formulated. In SM-ASRES, the optimization of multimedia encoding parameters is focused toward maximizing the number of served subscribers based only on user capability and channel conditions. On the other hand, RM-ASRES additionally considers subscribers' priority for multimedia broadcast adaptation. Energy saving at the user equipment is a result of time slicing approach at the transmission stage. As compared to the conventional broadcast schemes, the proposed user adaptive framework shows an appreciable improvement in quality of user experience and increased energy saving for hand-held mobile users, along with an increased revenue when employing RM-ASRES.

The reduction of energy consumption is a major concern in the current telecommunications environment - especially with the growth in usage of energy-hungry multimedia-centric applications on high-end mobile devices. In this context, this paper proposes eSMART, an Energy-efficient Scalable Multimedia Broadcast Transmission mechanism, that considers the energy-quality trade-off to reduce battery power consumption (increase energy saving) of heterogeneous mobile devices while maintaining acceptable perceived quality levels of received video. A real experimental test-bed has been built to analyze the impact of different multimedia scalability factors on the energy consumption of various mobile devices receiving broadcast content. Overall mobile device energy-saving is modeled using the accumulative effect of adaptive scalable video playback energy saving and time-sliced broadcast reception based radio-receiver's energy saving. eSMART's optimization framework performs user-centric adaptive encoding of scalable video that is broadcast to heterogeneous user equipments. eSMART serves more users at improved quality of experience levels and achieves up to 69% increase in mobile device energy savings as compared to a non-adaptive time-slicing scheme from the literature.

This paper presents a novel cross-layer optimization framework to improve the quality of user experience (QoE) and energy efficiency of the heterogeneous wireless multimedia broadcast receivers. This joint optimization is achieved by grouping the users based on their device capabilities and estimated channel conditions experienced by them and broadcasting adaptive content to these groups. The adaptive multimedia content is obtained by using scalable video coding (SVC) with optimal source encoding parameters resulted from an innovative cooperative game. Energy saving at user terminals results from using a layer-aware time slicing approach in the transmission stage. A trade-off between energy saving and QoE is observed, and is incorporated in the definition of a utility function of the players in the formulated heterogeneous user composition and physical channel aware game. An adaptive modulation and coding scheme is also optimally incorporated in order to maximize the reception quality of the broadcast receivers, while maximizing the network broadcast capacity. Compared to the conventional broadcast schemes, the proposed framework shows an appreciable improvement in QoE levels for all users, while achieving higher energy-savings for the energy constrained users.

In this paper, we present a new resource allocation scheme for cell-edge active users to achieve improved performance in terms of a higher system capacity and better quality-of-service (QoS) guarantee of the users, where we utilize the two-dimensional resource allocation flexibility of orthogonal frequency division multiple access (OFDMA) networks. Here, the mobile stations (MSs) at the cell-edge can maintain parallel connections with more than one base station (BS) when it is in their coverage area. A MS, before handoff to a new BS, seeks to utilize additional resources from the other BSs if the BS through which its current session is registered is not able to satisfy its requirements. The handoff procedure is termed as split handoff. The BSs participate in split handoff operation while guaranteeing that they are able to maintain QoS of the existing connections associated with them. In this study, first, we present the proposed shared resource allocation architecture and protocol functionalities in split handoff, and give a theoretical proof of concept of system capacity gain associated with the shared resource allocation approach. Then, we provide a differentiated QoS provisioning approach that accounts for the MS speed, its channel quality, as well as the loads at different BSs. Via extensive simulations in Qualnet, the benefits of the proposed class-based split handoff approach is demonstrated. The results also indicate traffic load balancing property of the proposed scheme in heavy traffic conditions.