This paper studies the computational offloading of CNN inference in device-edge co-inference systems. Inspired by the emerging paradigm semantic communication, we propose a novel autoencoder-based CNN architecture (AECNN), for effective feature extraction at end-device. We design a feature compression module based on the channel attention method in CNN, to compress the intermediate data by selecting the most important features. To further reduce communication overhead, we can use entropy encoding to remove the statistical redundancy in the compressed data. At the receiver, we design a lightweight decoder to reconstruct the intermediate data through learning from the received compressed data to improve accuracy. To fasten the convergence, we use a step-by-step approach to train the neural networks obtained based on ResNet-50 architecture. Experimental results show that AECNN can compress the intermediate data by more than 256x with only about 4% accuracy loss, which outperforms the state-of-the-art work, BottleNet++. Compared to offloading inference task directly to edge server, AECNN can complete inference task earlier, in particular, under poor wireless channel condition, which highlights the effectiveness of AECNN in guaranteeing higher accuracy within time constraint.
For time-critical IoT applications using deep learning, inference acceleration through distributed computing is a promising approach to meet a stringent deadline. In this paper, we implement a working prototype of a new distributed inference acceleration method HALP using three raspberry Pi 4. HALP accelerates inference by designing a seamless collaboration among edge devices (EDs) in Edge Computing. We maximize the parallelization between communication and computation among the collaborative EDs by optimizing the task partitioning ratio based on the segment-based partitioning. Experimental results show that the distributed inference HALP achieves 1.7x inference acceleration for VGG-16. Then, we combine distributed inference with conventional neural network model compression by setting up different shrinking hyperparameters for MobileNet-V1. In this way, we can further accelerate inference but at the cost of inference accuracy loss. To strike a balance between latency and accuracy, we propose dynamic model selection to select a model which provides the highest accuracy within the latency constraint. It is shown that the model selection with distributed inference HALP can significantly improve service reliability compared to the conventional stand-alone computation.
Wireless traffic is exploding, due to the myriad of new connections and the exchange of capillary data at the edge of the networks to operate real-time processing and decision making. The latter especially affects the uplink traffic, which will grow in 6G and beyond networks, calling for new optimization metrics that include energy, service delay, and electromagnetic field (EMF) exposure (EMFE). To this end, reconfigurable intelligent surfaces (RISs) represent a promising solution to mitigate the EMFE, thanks to their ability of shaping and manipulating the impinging electromagnetic waves. In line with this vision, this paper proposes an online adaptive method to mitigate the EMFE under end-to-end delay constraints of a computation offloading service, in the context of RIS and multi-access edge computing (MEC)-aided wireless networks. The goal is to minimize the long-term average of the EMF human exposure under such constraints, investigating the advantages of RISs towards blue (i.e. low EMFE) communications. A multiple-input multiple-output (MIMO) system is investigated as part of the visions towards 6G. Focusing on a typical scenario of computation offloading, the method jointly and adaptively optimizes user precoding, transmit power, RIS reflectivity parameters, and receiver combiner, with theoretical guarantees on the desired long-term performance. Besides the theoretical results, numerical simulations assess the performance of the proposed algorithm, when exploiting accurate antenna patterns, thus showing the advantage of the RIS and that of our method, compared to benchmark solutions.
Purpose: Visual perception enables robots to perceive the environment. Visual data is processed using computer vision algorithms that are usually time-expensive and require powerful devices to process the visual data in real-time, which is unfeasible for open-field robots with limited energy. This work benchmarks the performance of different heterogeneous platforms for object detection in real-time. This research benchmarks three architectures: embedded GPU -- Graphical Processing Units (such as NVIDIA Jetson Nano 2 GB and 4 GB, and NVIDIA Jetson TX2), TPU -- Tensor Processing Unit (such as Coral Dev Board TPU), and DPU -- Deep Learning Processor Unit (such as in AMD-Xilinx ZCU104 Development Board, and AMD-Xilinx Kria KV260 Starter Kit). Method: The authors used the RetinaNet ResNet-50 fine-tuned using the natural VineSet dataset. After the trained model was converted and compiled for target-specific hardware formats to improve the execution efficiency. Conclusions and Results: The platforms were assessed in terms of performance of the evaluation metrics and efficiency (time of inference). Graphical Processing Units (GPUs) were the slowest devices, running at 3 FPS to 5 FPS, and Field Programmable Gate Arrays (FPGAs) were the fastest devices, running at 14 FPS to 25 FPS. The efficiency of the Tensor Processing Unit (TPU) is irrelevant and similar to NVIDIA Jetson TX2. TPU and GPU are the most power-efficient, consuming about 5W. The performance differences, in the evaluation metrics, across devices are irrelevant and have an F1 of about 70 % and mean Average Precision (mAP) of about 60 %.
The Metaverse has emerged to extend our lifestyle beyond physical limitations. As essential components in the Metaverse, digital twins (DTs) are the real-time digital replicas of physical items. Multi-access edge computing (MEC) provides responsive services to the end users, ensuring an immersive and interactive Metaverse experience. While the digital representation (DT) of physical objects, end users, and edge computing systems is crucial in the Metaverse, the construction of these DTs and the interplay between them have not been well-investigated. In this paper, we discuss the bidirectional reliance between the DT and the MEC system and investigate the creation of DTs of objects and users on the MEC servers and DT-assisted edge computing (DTEC). To ensure seamless handover among MEC servers and to avoid intermittent Metaverse services, we also explore the interaction between local DTECs on local MEC servers and the global DTEC on the cloud server due to the dynamic nature of network states (e.g., channel state and users' mobility). We investigate a continual learning framework for resource allocation strategy in local DTEC through a case study. Our strategy mitigates the desynchronization between physical-digital twins, ensures higher learning outcomes, and provides a satisfactory Metaverse experience.
With the rise of edge computing, various AI services are expected to be available at a mobile side through the inference based on deep neural network (DNN) operated at the network edge, called edge inference (EI). On the other hand, the resulting AI quality (e.g., mean average precision in objective detection) has been regarded as a given factor, and AI quality control has yet to be explored despite its importance in addressing the diverse demands of different users. This work aims at tackling the issue by proposing a feature hierarchical EI (FHEI), comprising feature network and inference network deployed at an edge server and corresponding mobile, respectively. Specifically, feature network is designed based on feature hierarchy, a one-directional feature dependency with a different scale. A higher scale feature requires more computation and communication loads while it provides a better AI quality. The tradeoff enables FHEI to control AI quality gradually w.r.t. communication and computation loads, leading to deriving a near-to-optimal solution to maximize multi-user AI quality under the constraints of uplink \& downlink transmissions and edge server and mobile computation capabilities. It is verified by extensive simulations that the proposed joint communication-and-computation control on FHEI architecture always outperforms several benchmarks by differentiating each user's AI quality depending on the communication and computation conditions.
Federated Learning (FL) with quantization and deliberately added noise over wireless networks is a promising approach to preserve user differential privacy (DP) while reducing wireless resources. Specifically, an FL process can be fused with quantized Binomial mechanism-based updates contributed by multiple users. However, optimizing quantization parameters, communication resources (e.g., transmit power, bandwidth, and quantization bits), and the added noise to guarantee the DP requirement and performance of the learned FL model remains an open and challenging problem. This article aims to jointly optimize the quantization and Binomial mechanism parameters and communication resources to maximize the convergence rate under the constraints of the wireless network and DP requirement. To that end, we first derive a novel DP budget estimation of the FL with quantization/noise that is tighter than the state-of-the-art bound. We then provide a theoretical bound on the convergence rate. This theoretical bound is decomposed into two components, including the variance of the global gradient and the quadratic bias that can be minimized by optimizing the communication resources, and quantization/noise parameters. The resulting optimization turns out to be a Mixed-Integer Non-linear Programming (MINLP) problem. To tackle it, we first transform this MINLP problem into a new problem whose solutions are proved to be the optimal solutions of the original one. We then propose an approximate algorithm to solve the transformed problem with an arbitrary relative error guarantee. Extensive simulations show that under the same wireless resource constraints and DP protection requirements, the proposed approximate algorithm achieves an accuracy close to the accuracy of the conventional FL without quantization/noise. The results can achieve a higher convergence rate while preserving users' privacy.
Computer vision applications in intelligent transportation systems (ITS) and autonomous driving (AD) have gravitated towards deep neural network architectures in recent years. While performance seems to be improving on benchmark datasets, many real-world challenges are yet to be adequately considered in research. This paper conducted an extensive literature review on the applications of computer vision in ITS and AD, and discusses challenges related to data, models, and complex urban environments. The data challenges are associated with the collection and labeling of training data and its relevance to real world conditions, bias inherent in datasets, the high volume of data needed to be processed, and privacy concerns. Deep learning (DL) models are commonly too complex for real-time processing on embedded hardware, lack explainability and generalizability, and are hard to test in real-world settings. Complex urban traffic environments have irregular lighting and occlusions, and surveillance cameras can be mounted at a variety of angles, gather dirt, shake in the wind, while the traffic conditions are highly heterogeneous, with violation of rules and complex interactions in crowded scenarios. Some representative applications that suffer from these problems are traffic flow estimation, congestion detection, autonomous driving perception, vehicle interaction, and edge computing for practical deployment. The possible ways of dealing with the challenges are also explored while prioritizing practical deployment.
The Metaverse has emerged as the successor of the conventional mobile internet to change people's lifestyles. It has strict visual and physical requirements to ensure an immersive experience (i.e., high visual quality, low motion-to-photon latency, and real-time tactile and control experience). However, the current technologies fall short to satisfy these requirements. Mobile edge computing (MEC) has been indispensable to enable low latency and powerful computing. Moreover, the sixth generation (6G) networks promise to provide end users with seamless communications. In this paper, we explore and demonstrate the synergistic relationship between 6G and mobile-edge technologies in empowering the Metaverse with ubiquitous communications and computation. This includes the usage of heterogeneous radios, intelligent reflecting surfaces (IRS), non-orthogonal multiple access (NOMA), and digital twins (DTs) - assisted MEC. We also discuss emerging communication paradigms (i.e., semantic communication, holographic-type communication, and haptic communication) to further satisfy the demand for human-type communications and fulfill user preferences and immersive experiences in the Metaverse.
As the Internet grows in popularity, more and more classification jobs, such as IoT, finance industry and healthcare field, rely on mobile edge computing to advance machine learning. In the medical industry, however, good diagnostic accuracy necessitates the combination of large amounts of labeled data to train the model, which is difficult and expensive to collect and risks jeopardizing patients' privacy. In this paper, we offer a novel medical diagnostic framework that employs a federated learning platform to ensure patient data privacy by transferring classification algorithms acquired in a labeled domain to a domain with sparse or missing labeled data. Rather than using a generative adversarial network, our framework uses a discriminative model to build multiple classification loss functions with the goal of improving diagnostic accuracy. It also avoids the difficulty of collecting large amounts of labeled data or the high cost of generating large amount of sample data. Experiments on real-world image datasets demonstrates that the suggested adversarial federated transfer learning method is promising for real-world medical diagnosis applications that use image classification.
Owing to the widespread adoption of the Internet of Things, a vast amount of sensor information is being acquired in real time. Accordingly, the communication cost of data from edge devices is increasing. Compressed sensing (CS), a data compression method that can be used on edge devices, has been attracting attention as a method to reduce communication costs. In CS, estimating the appropriate compression ratio is important. There is a method to adaptively estimate the compression ratio for the acquired data using reinforcement learning (RL). However, the computational costs associated with existing RL methods that can be utilized on edges are often high. In this study, we developed an efficient RL method for edge devices, referred to as the actor--critic online sequential extreme learning machine (AC-OSELM), and a system to compress data by estimating an appropriate compression ratio on the edge using AC-OSELM. The performance of the proposed method in estimating the compression ratio is evaluated by comparing it with other RL methods for edge devices. The experimental results indicate that AC-OSELM demonstrated the same or better compression performance and faster compression ratio estimation than the existing methods.
In this current technological world, the application of machine learning is becoming ubiquitous. Incorporating machine learning algorithms on extremely low-power and inexpensive embedded devices at the edge level is now possible due to the combination of the Internet of Things (IoT) and edge computing. To estimate an outcome, traditional machine learning demands vast amounts of resources. The TinyML concept for embedded machine learning attempts to push such diversity from usual high-end approaches to low-end applications. TinyML is a rapidly expanding interdisciplinary topic at the convergence of machine learning, software, and hardware centered on deploying deep neural network models on embedded (micro-controller-driven) systems. TinyML will pave the way for novel edge-level services and applications that survive on distributed edge inferring and independent decision-making rather than server computation. In this paper, we explore TinyML's methodology, how TinyML can benefit a few specific industrial fields, its obstacles, and its future scope.
Edge computing-based 3D perception has received attention in intelligent transportation systems (ITS) because real-time monitoring of traffic candidates potentially strengthens Vehicle-to-Everything (V2X) orchestration. Thanks to the capability of precisely measuring the depth information on surroundings from LiDAR, the increasing studies focus on lidar-based 3D detection, which significantly promotes the development of 3D perception. Few methods met the real-time requirement of edge deployment because of high computation-intensive operations. Moreover, an inconsistency problem of object detection remains uncovered in the pointcloud domain due to large sparsity. This paper thoroughly analyses this problem, comprehensively roused by recent works on determining inconsistency problems in the image specialisation. Therefore, we proposed a 3D harmonic loss function to relieve the pointcloud based inconsistent predictions. Moreover, the feasibility of 3D harmonic loss is demonstrated from a mathematical optimization perspective. The KITTI dataset and DAIR-V2X-I dataset are used for simulations, and our proposed method considerably improves the performance than benchmark models. Further, the simulative deployment on an edge device (Jetson Xavier TX) validates our proposed model's efficiency.
Event-based neuromorphic systems provide a low-power solution by using artificial neurons and synapses to process data asynchronously in the form of spikes. Ferroelectric Tunnel Junctions (FTJs) are ultra low-power memory devices and are well-suited to be integrated in these systems. Here, we present a hybrid FTJ-CMOS Integrate-and-Fire neuron which constitutes a fundamental building block for new-generation neuromorphic networks for edge computing. We demonstrate electrically tunable neural dynamics achievable by tuning the switching of the FTJ device.
Federated learning enables a large amount of edge computing devices to learn a model without data sharing jointly. As a leading algorithm in this setting, Federated Average FedAvg, which runs Stochastic Gradient Descent (SGD) in parallel on local devices and averages the sequences only once in a while, have been widely used due to their simplicity and low communication cost. However, despite recent research efforts, it lacks theoretical analysis under assumptions beyond smoothness. In this paper, we analyze the convergence of FedAvg. Different from the existing work, we relax the assumption of strong smoothness. More specifically, we assume the semi-smoothness and semi-Lipschitz properties for the loss function, which have an additional first-order term in assumption definitions. In addition, we also assume bound on the gradient, which is weaker than the commonly used bounded gradient assumption in the convergence analysis scheme. As a solution, this paper provides a theoretical convergence study on Federated Learning.
5G wireless technology can deliver higher data speeds, ultra low latency, more reliability, massive network capacity, increased availability, and a more uniform user experience to users. It brings additional power to help address the challenges brought by renewable integration and decarbonization. In this paper, a 5G enabled adaptive computing workflow has been presented that consists of various computing resources, such as 5G equipment, edge computing, cluster, Graphics processing unit (GPU) and cloud computing, with two examples showing technical feasibility for edge-grid-cloud interaction for power system real-time monitoring, security assessment, and forecasting. Benefiting from the high speed data transport and massive connection capability of 5G, the workflow shows its potential to seamlessly integrate various applications at distributed and/or centralized locations to build more complex and powerful functions, with better flexibility.
The recent drastic increase in mobile data traffic has pushed the mobile edge computing systems to the limit of their capacity. A promising solution to this problem is the task migration provided by unmanned aerial vehicles (UAV). Key factors to be taken into account in the design of UAV offloading schemes must include the number of tasks waiting in the system as well as their corresponding deadlines. An appropriate system cost which is used as an objective function to be minimized comprises two parts. First, an offloading cost which can be interpreted as the cost of using computational resources at the UAV. Second, a penalty cost due to potential task expiration. In order to minimize the expected (time average) cost over a time horizon, we formulate a Dynamic Programming (DP) equation and analyze it to describe properties of a candidate optimal offloading policy. The DP equation suffers from the well-known "Curse of Dimensionality" that makes computations intractable, especially when the state space is infinite. In order to reduce the computational burden, we identify three important properties of the optimal policy. Based on these properties, we show that it suffices to evaluate the DP equation on a finite subset of the state space only. We then show that the optimal task offloading decision associated with a state can be inferred from the decision taken at its "adjacent" states, further reducing the computational load. Finally, we provide numerical results to evaluate the influence of different parameters on the system performance as well as verify the theoretical results.
Mobile Edge Computing (MEC) has emerged as an alternative to cloud computing to meet the latency and Quality-of-Service (QoS) requirements of mobile devices. In this paper, we address the problem of server resource allocation in MEC. Due to the dynamic load conditions on MEC servers, their resources need to be used intelligently to meet the QoS requirements of the users and to minimize server energy consumption. We present a novel resource allocation algorithm, called Power Migration Expand (PowMigExpand). Our algorithm assigns user requests to the optimal server and allocates optimal amount of resources to User Equipment (UE) based on our comprehensive utility function. PowMigExpand also migrates UE requests to new servers, when needed due to the mobility of users. We also present a low cost Energy Efficient Smart Allocator (EESA) algorithm that uses deep learning for energy efficient allocation of requests to optimal servers. The proposed algorithms consider varying load of incoming requests and their heterogeneous nature, energy efficient activation of servers, and Virtual Machine (VM) migration for smart resource allocation and, thus, is the first comprehensive approach to address the complex and multidimensional resource allocation problem using deep learning. We compare our proposed algorithms with other resource allocation approaches and show that our approach can handle the dynamic load conditions better. The proposed algorithms improve the service rate and the overall utility with minimum energy consumption. On average, it reduces 26% energy consumption of MESs and improves the service rate by 23%, compared with other algorithms. We also get more than 70% accuracy for EESA in allocating the resources of multiple servers to multiple users.
The anti-interference capability of wireless links is a physical layer problem for edge computing. Although convolutional codes have inherent error correction potential due to the redundancy introduced in the data, the performance of the convolutional code is drastically degraded due to multipath effects on the channel. In this paper, we propose the use of a Hidden Markov Model (HMM) for the reconstruction of convolutional codes and decoding by the Viterbi algorithm. Furthermore, to implement soft-decision decoding, the observation of HMM is replaced by Gaussian mixture models (GMM). Our method provides superior error correction potential than the standard method because the model parameters contain channel state information (CSI). We evaluated the performance of the method compared to standard Viterbi decoding by numerical simulation. In the multipath channel, the hybrid HMM decoder can achieve a performance gain of 4.7 dB and 2 dB when using hard-decision and soft-decision decoding, respectively. The HMM decoder also achieves significant performance gains for the RSC code, suggesting that the method could be extended to turbo codes.
Deep convolutional neural networks (DCNNs) have become the state-of-the-art (SOTA) approach for many computer vision tasks: image classification, object detection, semantic segmentation, etc. However, most SOTA networks are too large for edge computing. Here, we suggest a simple way to reduce the number of trainable parameters and thus the memory footprint: sharing kernels between multiple convolutional layers. Kernel-sharing is only possible between ``isomorphic" layers, i.e.layers having the same kernel size, input and output channels. This is typically the case inside each stage of a DCNN. Our experiments on CIFAR-10 and CIFAR-100, using the ConvMixer and SE-ResNet architectures show that the number of parameters of these models can drastically be reduced with minimal cost on accuracy. The resulting networks are appealing for certain edge computing applications that are subject to severe memory constraints, and even more interesting if leveraging "frozen weights" hardware accelerators. Kernel-sharing is also an efficient regularization method, which can reduce overfitting. The codes are publicly available at https://github.com/AlirezaAzadbakht/kernel-sharing.