J. Sens. Actuator Netw., Volume 9, Issue 4 (December 2020) – 15 articles

During the last decade, the Internet of Things acted as catalyst for the big data phenomenon. As result, modern edge devices can access a huge amount of data that can be exploited to build useful services. In such a context, artificial intelligence has a key role to develop intelligent systems (e.g., intelligent cyber physical systems) that create a connecting bridge with the physical world. However, as time goes by, machine and deep learning applications are becoming more complex, requiring increasing amounts of data and training time, which makes the use of centralized approaches unsuitable. Federated learning is an emerging paradigm which enables the cooperation of edge devices to learn a shared model (while keeping private their training data), thereby abating the training time. Although federated learning is a promising technique, its implementation is difficult and brings a lot of challenges. In this paper, we present an extension of Stack4Things, a cloud platform developed in our department; leveraging its functionalities, we enabled the deployment of federated learning on edge devices without caring their heterogeneity. Experimental results show a comparison with a centralized approach and demonstrate the effectiveness of the proposed approach in terms of both training time and model accuracy. Full article

5G systems are putting increasing pressure on Telecom operators to enhance users’ experience, leading to the development of more techniques with the aim of improving service quality. However, it is essential to take into consideration not only users’ demands but also service providers’ interests. In this work, we explore policies that satisfy both views. We first formulate a mathematical model to compute End-to-End (E2E) delay experienced by mobile users in Multi-access Edge Computing (MEC) environments. Then, dynamic Virtual Machine (VM) allocation policies are presented, with the objective of satisfying mobile users Quality of Service (QoS) requirements, while optimally using the cloud resources by exploiting VM resource reuse.Thus, maximizing the service providers’ profit should be ensured while providing the service required by users. We further demonstrate the benefits of these policies in comparison with previous works. Full article

Multi-access edge computing (MEC) promises to deliver localized computing power and storage. Coupled with low-latency 5G radio access, this enables the creation of high added-value services for mobile users, such as in-vehicle infotainment or remote driving. The performance of these services as well as their scalability will however depend on how MEC will be deployed in 5G systems. This paper evaluates different MEC deployment options, coherent with the respective 5G migration phases, using an accurate and comprehensive end-to-end (E2E) system simulation model (exploiting Simu5G for radio access and Intel CoFluent for core network and MEC), taking into account user-related metrics, such as response time or MEC latency. Our results show that 4G radio access is going to be a bottleneck, preventing MEC services from scaling up. On the other hand, the introduction of 5G will allow a considerable higher penetration of MEC services. Full article

Autonomous systems developed with the Belief-Desire-Intention (BDI) architecture tend to be mostly implemented in simulated environments. In this project we sought to build a BDI agent for use in the real world for campus mail delivery in the tunnel system at Carleton University. Ideally, the robot should receive a delivery order via a mobile application, pick up the mail at a station, navigate the tunnels to the destination station, and notify the recipient. In this paper, we discuss how we linked the Robot Operating System (ROS) with a BDI reasoning system to achieve a subset of the required use casesand demonstrated the system performance in an analogue environment. ROS handles the connections to the low-level sensors and actuators, while the BDI reasoning system handles the high-level reasoning and decision making. Sensory data is sent to the reasoning system as perceptions using ROS. These perceptions are then deliberated upon, and an action string is sent back to ROS for interpretation and driving of the necessary actuator for the action to be performed. In this paper we present our current implementation, which closes the loop on the hardware-software integration and implements a subset of the use cases required for the full system. We demonstrated the performance of the system in an analogue environment. Full article

With reference to the MeSmart project, the Municipality of Messina is making a great investments to deploy several types of cameras and digital devices across the city for carrying out different tasks related to mobility management, such as traffic flow monitoring, number plate recognition, video surveillance etc. To this aim, exploiting specific devices for each task increases infrastructure and management costs, reducing flexibility. On the contrary, using general-purpose devices customized to accomplish multiple tasks at the same time can be a more efficient solution. Another important approach that can improve the efficiency of mobility services is moving computation tasks at the Edge of the managed system instead of in remote centralized serves, so reducing delays in event detection and processing and making the system more scalable. In this paper, we investigate the adoption of Edge devices connected to high-resolution cameras to create a general-purpose solution for performing different tasks. In particular, we use the Function as a Service (FaaS) paradigm to easily configure the Edge device and set up new services. The key results of our work is deploying versatile, scalable and adaptable systems able to respond to smart city’s needs, even if such needs change over time. We tested the proposed solution setting up a vehicle counting solution based on OpenCV, and automatically deploying necessary functions into an Edge device. From experimental results, it results that computing performance at the Edge overtakes the performance of a device specifically designed for vehicle counting under certain conditions and thanks to our reconfigurable functions. Full article

Face recognition (FR) in an unconstrained environment, such as low light, illumination variations, and bad weather is very challenging and still needs intensive further study. Previously, numerous experiments on FR in an unconstrained environment have been assessed using Eigenface, Fisherface, and Local binary pattern histogram (LBPH) algorithms. The result indicates that LBPH FR is the optimal one compared to others due to its robustness in various lighting conditions. However, no specific experiment has been conducted to identify the best setting of four parameters of LBPH, radius, neighbors, grid, and the threshold value, for FR techniques in terms of accuracy and computation time. Additionally, the overall performance of LBPH in the unconstrained environments are usually underestimated. Therefore, in this work, an in-depth experiment is carried out to evaluate the four LBPH parameters using two face datasets: Lamar University data base (LUDB) and 5_celebrity dataset, and a novel Bilateral Median Convolution-Local binary pattern histogram (BMC-LBPH) method was proposed and examined in real-time in rainy weather using an unmanned aerial vehicle (UAV) incorporates with 4 vision sensors. The experimental results showed that the proposed BMC-LBPH FR techniques outperformed the traditional LBPH methods by achieving the accuracy of $65\%$, $98\%$, and $78\%$ in 5_celebrity dataset, LU dataset, and rainy weather, respectively. Ultimately, the proposed method provides a promising solution for facial recognition using UAV. Full article

Mobile network traffic is increasing in an unprecedented manner, resulting in growing demand from network operators to deploy more base stations able to serve more devices while maintaining a satisfactory level of service quality. Base stations are considered the leading energy consumer in network infrastructure; consequently, increasing the number of base stations will increase power consumption. By predicting the traffic load on base stations, network optimization techniques can be applied to decrease energy consumption. This research explores different machine learning and statistical methods capable of predicting traffic load on base stations. These methods are examined on a public dataset that provides records of traffic loads of several base stations over the span of one week. Because of the limited number of records in the dataset for each base station, different base stations are grouped while building the prediction model. Due to the different behavior of the base stations, forecasting the traffic load of multiple base stations together becomes challenging. The proposed solution involves clustering the base stations according to their behavior and forecasting the load on the base stations in each cluster individually. Clustering the time series data according to their behavior mitigates the dissimilar behavior problem of the time series when they are trained together. Our findings demonstrate that predictions based on deep recurrent neural networks perform better than other forecasting techniques. Full article

One of the current topics of interest in transportation science is the use of intelligent computation and IoT (Internet of Things) technologies. Researchers have proposed many approaches using these concepts, but the most widely used concept in road traffic modeling at the microscopic level is the car-following model. Knowing that the standard car-following model is single lane-oriented, the purpose of this paper is to present a fault detection analysis of the extension to a multiple lane car-following model that uses the Bayesian reasoning concept to estimate lane change behavior. After the application of the latter model on real traffic data retrieved from inductive loops placed on a road network, fault detection using parity equations was used. The standard car-following model applied separately for each lane showed the ability to perform a lane change action and to incorporate a new vehicle into the current lane. The results will highlight the advantages and the critical points of influence in the use of a multiple lane car-following model based on probabilistic estimated lane changes. Additionally, this research applied fault detection based on parity equations for the proposed model. The purpose was to deliver an overview of the faults introduced by the behavior of vehicles in adjacent lanes on the behavior of the target vehicle. Full article

In this paper, we propose a single chip fingerprint sensor with the algorithm processor and 16-bit MCU. The algorithm processor is a logic circuit that implements the GABOR filter and the THINNING step, which occupies 80% of the fingerprint image processing time. The rest of the algorithm is processed by embedded 16-bit MCU with small circuit volume, so all steps of the algorithm can be processed on the single chip without an external CPU. The capacitive sensing circuit was designed by applying the parasitic-insensitive integrator with the variable clock generator. The function was verified by Cadence Spectre for a 1-pixel sensor scheme and RTL and post simulation for digital blocks synthesized by Synopsys Design Compiler in 180n 2-poly 6-metal CMOS (complementary metal–oxide–semiconductor) process. The layout is done by automatic P&R for the full chip in a 96 × 96 pixel array. The chip area is 5010 μm × 5710 μm (28.61 mm²) and the gate count is 2,866,700. The result is compared with a conventional one. The proposed scheme can reduce the processing time by 57%. Full article

The need to protect sensitive data is growing, and environmental data are now considered sensitive. The application of last-generation procedures such as blockchains coupled with the implementation of new air quality monitoring technology allows the data protection and validation. In this work, the use of a blockchain applied to air pollution data is proposed. A blockchain procedure has been designed and tested. An Internet of Things (IoT)-based sensor network provides air quality data in terms of particulate matter of two different diameters, particulate matter (PM)₁₀ and PM_2.5, volatile organic compounds (VOC), and nitrogen dioxide (NO₂) concentrations. The dataset also includes meteorological parameters and vehicular traffic information. This work foresees that the data, recovered from traditional Not Structured Query Language (NoSQL) database, and organized according to some specifications, are sent to the Ethereum blockchain daily automatically and with the possibility to choose the period of interest manually. There was also the development of a transaction management and recovery system aimed at retrieving data, formatting it according to the specifications and organizing it into files of various formats. The blockchain procedure has therefore been used to track data provided by air quality monitoring networks unequivocally. Full article

Heterogeneous networks (HetNets), which are combined with a macrocell and picocell in the same coverage, are expected to further increase the system capacity in fifth-generation mobile systems and beyond. In HetNets, the cell range expansion (CRE) technique plays an important role and can allow more user equipment (UE) to access the picocell, i.e., virtually expand the picocell coverage. However, conventional CRE techniques that provide a fixed cell selection offset (CSO) for all UE may worsen user throughput if UE is forced to connect to the picocell because the received signal-to-interference plus noise ratio of the UE becomes lower. Therefore, we propose a personal picocell scheme using an adaptive control CRE technique to improve user throughput in which different CSOs are assigned to UE to form each optimal picocell for each UE. In this paper, we first describe the aspects and algorithm of the proposed scheme. Then, we show the user throughput for adaptive control CRE in comparison with conventional CRE by using system-level computer simulations for the two types of HetNets, i.e., single-band and multi-band HetNets. In the simulations, we first clarify the optimal parameters of the adaptive control CRE. We then show the average and 5-percentile user throughput of the optimized adaptive control CRE in comparison with that of conventional CRE. From these results, we confirmed that the personal picocell scheme using the adaptive control CRE can improve the 5-percentile user throughput while maintaining the average user throughput compared with that of conventional CRE. Full article

The adoption of emergency alerting systems can bring countless benefits when managing urban areas, industrial plants, farms, roads and virtually any area that is subject to the occurrence of critical events, supporting in rescue operations and reducing their negative impacts. For such systems, a promising approach is to exploit scalar sensors to detect events of interest, allowing for the distributed monitoring of different variables. However, the use of cameras as visual sensors can enhance the detection of critical events, which can be employed along with scalar sensors for a more comprehensive perception of the environment. Although the particularities of visual sensing may be challenging in some scenarios, the combination of scalar and visual sensors for the early detection of emergency situations can be valuable for many scenarios, such as smart cities and industry 4.0, bringing promising results. Therefore, in this article, we extend a sensors-based emergency detection and alerting system to also exploit visual monitoring when identifying critical events. Implementation and experimental details are provided to reinforce the use of cameras as a relevant sensor unit, bringing promising results for emergencies management. Full article

Wireless Sensor Network (WSN) software development challenges developers in two main ways: through system programming, which requires expertise in hardware and network management; and application programming, which requires domain-specific knowledge. However, domain programmers often lack WSN programming expertise. Likewise, system-specific programmers may find it difficult to understand domain-specific requirements. As a result, domain programmers often refrain from using WSN technology in domain-specific applications. Therefore, we propose a Finite State Machine (FSM)-based approach with an affiliated framework to decouple application functionality from WSN details. Instead of the traditional flat FSM, we use statecharts formalism because of its relaxed definition of system states. In this paper, we compare the statecharts paradigm against two basic WSN sensor node programming frameworks. The result exhibits that statecharts are an advanced paradigm in WSN application development. It motivated us to develop a statecharts framework. In our framework, we choose not to use the typical solution which converts statecharts to programming code. Instead of that, we implement a statecharts middleware associated with action libraries to interpret and actuate raw statecharts on an operating system. This approach allows domain programmers to concentrate on WSN application behavior, and system-specific programmers to focus on developing WSN services. We also introduce our statecharts middleware and present a living example with performance evaluation. Full article