With the ever-increasing popularity of wearable devices, data on the time and location of popular walking, running, and bicycling routes is expansive and growing rapidly. These data are currently ...used primarily for route discovery and mobile context awareness, as it provides precise and updated information about urban dynamics. We leverage these data to build ad hoc transportation flows, and we present a novel model that creates delivery networks from these zero-emission transportation flows. We evaluate the model using data from two popular datasets, and our results indicate that such networks are indeed possible, and can help reduce traffic, emissions, and delivery times. Moreover, we demonstrate how our results can be consistently reproduced in different cities with different subsets of carriers. We then extend our work into predicting routes of vehicles, hence possible delivery flows, based on the traces history. We conclude this paper by laying the groundwork for a future real-world study.
As the complexity of Deep Neural Network (DNN) models increases, their deployment on mobile devices becomes increasingly challenging, especially in complex vision tasks such as image classification. ...Many of recent contributions aim either to produce compact models matching the limited computing capabilities of mobile devices or to offload the execution of such burdensome models to a compute-capable device at the network edge - the edge servers. In this paper, we propose to modify the structure and training process of DNN models for complex image classification tasks to achieve in-network compression in the early network layers. Our training process stems from knowledge distillation, a technique that has been traditionally used to build small - student - models mimicking the output of larger - teacher - models. Here, we adopt this idea to obtain aggressive compression while preserving accuracy. Our results demonstrate that our approach is effective for state-of-the-art models trained over complex datasets, and can extend the parameter region in which edge computing is a viable and advantageous option. Additionally, we demonstrate that in many settings of practical interest we reduce the inference time with respect to specialized models such as MobileNet v2 executed at the mobile device, while improving accuracy.
The main goal of an Indoor Positioning System (IPS) is to estimate the position of mobile devices in indoor environments. For this purpose, the primary source of information is the signal strength of ...packets received by a set of routers. The fingerprint technique is one of the most used techniques for IPSs. By using supervised machine learning techniques, it trains a model with the received signal intensity information so it can be used to estimate the positions of the devices later in an online phase. Although the k-Nearest Neighbors (kNN) is one of the most widely used classification methods due to its accuracy, it has no scalability since a sample that needs to be classified must be compared to all other samples in the training database. In this work, we use a novel hierarchical navigable small world graph technique to build a search structure so the location of a sample can be efficiently found, allowing the IPSs to be used in large-scale scenarios or run on devices with limited resources. To carry out our performance evaluation, we proposed a synthetic IPS dataset generator as well as implemented a complete real-world, large-scale IPS testbed. We compared the performance of our graph-based solution with other known kNN variants, such as Kd-Tree and Ball-Tree. Our results clearly show the performance gains of the proposed solution at 98% when compared to the classic kNN and at least 80% when compared to tree-based approaches.
The ability of Unmanned Aerial Vehicles (UAV) to autonomously operate is constrained by the severe limitations of their on-board resources. The limited processing capacity and energy storage of these ...devices inevitably makes the real-time analysis of complex signals - the key to autonomy - challenging. In urban environments, the UAVs can leverage the communication and computation resources of the surrounding city-wide Internet of Things infrastructure to enhance their capabilities. For instance, the UAVs can interconnect with edge computing resources and offload computation tasks to improve response time to sensor input and reduce energy consumption. However, the complexity of the urban topology and large number of devices and data streams competing for the same network and computation resources create an extremely dynamic environment, where poor channel conditions and edge server congestion may penalize the performance of task offloading. This paper develops a framework enabling optimal offloading decisions as a function of network and computation load parameters and current state. The optimization is formulated as an optimal stopping time problem over a semi-Markov process. We solve the optimization problem using Dynamic Programming and Deep Reinforcement learning at different levels of abstraction and prior knowledge of the system underlying stochastic processes. We validate our results in a realistic scenario, where a UAV performs a building inspection task while connected to an edge server.
We propose a low-latency approach for generating secure electrocardiogram (ECG) feature-based cryptographic keys. This is done by taking advantage of the uniqueness and randomness properties of ECG's ...main features. This approach achieves a low-latency since the key generation relies on four reference-free ECG's main features that can be acquired in short time. We call the approach several ECG features (SEF)-based cryptographic key generation. SEF consists of: 1) detecting the arrival time of ECG's fiducial points using Daubechies wavelet transform to compute ECG's main features accordingly; 2) using a dynamic technique to specify the optimum number of bits that can be extracted from each main ECG feature, comprising of PR, RR, PP, QT, and ST intervals; 3) generating cryptographic keys by exploiting the above-mentioned ECG features; and 4) consolidating and strengthening the SEF approach with cryptographically secure pseudo-random number generators. Fibonacci linear feedback shift register and advanced encryption standard algorithms are implemented as the pseudo-random number generator to enhance the security level of the generated cryptographic keys. Our approach is applied to 239 subjects' ECG signals comprising of normal sinus rhythm, arrhythmia, atrial fibrillation, and myocardial infraction. The security analyses of the proposed approach are carried out in terms of distinctiveness, test of randomness, temporal variance, and using National Institute of Standards and Technology benchmark. The analyses reveal that the normal ECG rhythms have slightly better randomness compared with the abnormal ones. The analyses also show that the strengthened SEF key generation approach provides a higher security level in comparison to existing approaches that rely only on singleton ECG features. For the normal ECG rhythms, the SEF approach has in average the entropy of about 0.98 while cryptographic keys which are generated utilizing the strengthened SEF approach offer the entropy of ~1. The execution time required to generate the cryptographic keys on different processors is also examined. The results reveal that our SEF approach is in average 1.8 times faster than existing key generation approaches which only utilize the inter pulse interval feature of ECG.
State-of-the-art drone technologies have severe flight time limitations due to weight constraints, which inevitably lead to a relatively small amount of available energy. Therefore, frequent battery ...replacement or recharging is necessary in applications such as delivery, exploration, or support to the wireless infrastructure. Mobile charging stations (i.e., mobile stations with charging equipment) for outdoor ad-hoc battery charging is one of the feasible solutions to address this issue. However, the ability of these platforms to charge the drones is limited in terms of the number and charging time. This paper designs an auction-based mechanism to control the charging schedule in multi-drone setting. In this paper, charging time slots are auctioned, and their assignment is determined by a bidding process. The main challenge in developing this framework is the lack of prior knowledge on the distribution of the number of drones participating in the auction. Based on optimal second-price-auction, the proposed formulation, then, relies on deep learning algorithms to learn such distribution online. Numerical results from extensive simulations show that the proposed deep-learning-based approach provides effective battery charging control in multi-drone scenarios.
This paper presents a battery-aware stochastic control framework for residential energy management systems (EMS) equipped with energy harvesting, that is, photovoltaic panels, and storage ...capabilities. The model and control rationale takes into account the dynamics of load, the weather, the weather forecast, the utility, and consumer preferences into a unified Markov decision process. The embedded optimization problem is formulated to determine the proportion of energy drawn from the battery and the grid to minimize a cost function capturing a user-defined tradeoff between battery degradation and financial expense by user preferences. Numerical results are based on real-world weather data for Golden, Colorado, and load traces. The results illustrate the ability of the system to limit battery degradation assessed using the Rain flow counting method for lithium ion batteries.
The edge computing paradigm places compute-capable devices - edge servers - at the network edge to assist mobile devices in executing data analysis tasks. Intuitively, offloading compute-intense ...tasks to edge servers can reduce their execution time. However, poor conditions of the wireless channel connecting the mobile devices to the edge servers may degrade the overall capture-to-output delay achieved by edge offloading. Herein, we focus on edge computing supporting remote object detection by means of Deep Neural Networks (DNNs), and develop a framework to reduce the amount of data transmitted over the wireless link. The core idea we propose builds on recent approaches splitting DNNs into sections - namely head and tail models - executed by the mobile device and edge server, respectively. The wireless link, then, is used to transport the output of the last layer of the head model to the edge server, instead of the DNN input. Most prior work focuses on classification tasks and leaves the DNN structure unaltered. Herein, our focus is on DNNs for three different object detection tasks, which present a much more convoluted structure, and modify the architecture of the network to: (i) achieve in-network compression by introducing a bottleneck layer in the early layers on the head model, and (ii) prefilter pictures that do not contain objects of interest using a convolutional neural network. Results show that the proposed technique represents an effective intermediate option between local and edge computing in a parameter region where these extreme point solutions fail to provide satisfactory performance. The code and trained models are available at https://github.com/yoshitomo-matsubaralhnd-ghnd-object-detectors.
In this paper, we address the problem of packet forwarding in multi-hop wireless networks of nodes equipped with multiple antennas. Through the use of MIMO (multiple-input multiple-output) ...technology, the physical layer is able to superimpose multiple concurrent information flows, thereby increasing the communications parallelism in the network. We propose a cooperative cross-layer scheme that integrates distributed incremental redundancy hybrid automatic retransmission request (HARQ) error control with routing in order to increase the efficiency of the transmissions and decrease the interference load in the network. In the proposed scheme, next hop relays are opportunistically selected during the HARQ process in order to avoid overloaded receivers and guarantee the geographical advancement of the packets. Thus, cooperation is used to both strengthen weak links and select the path to the destination. Numerical results are presented for a specific instantiation where a layered space-time multiuser detection (LASTMUD) transmitter-receiver architecture with spread-spectrum signals is implemented at the physical layer and packets are encoded with a linear erasure code (LEC).