Convolutional neural networks (CNNs) have yielded state-of-the-art performance in image classification and other computer vision tasks. Their application in fire detection systems will substantially improve detection accuracy, which will eventually minimize fire disasters and reduce the ecological and social ramifications. However, the major concern with CNN-based fire detection systems is their implementation in real-world surveillance networks, due to their high memory and computational requirements for inference. In this paper, we propose an original, energy-friendly, and computationally efficient CNN architecture, inspired by the SqueezeNet architecture for fire detection, localization, and semantic understanding of the scene of the fire. It uses smaller convolutional kernels and contains no dense, fully connected layers, which helps keep the computational requirements to a minimum. Despite its low computational needs, the experimental results demonstrate that our proposed solution achieves accuracies that are comparable to other, more complex models, mainly due to its increased depth. Moreover, this paper shows how a tradeoff can be reached between fire detection accuracy and efficiency, by considering the specific characteristics of the problem of interest and the variety of fire data.

Ubiquitous smart environments, equipped with low-cost and easy-deployable wireless sensor networks (WSNs) and widespread mobile ad hoc networks (MANETs), are opening brand new opportunities in wide-scale urban monitoring. Indeed, MANET and WSN convergence paves the way for the development of brand new Internet of Things (IoT) communication platforms with a high potential for a wide range of applications in different domains. Urban data collection, i.e., the harvesting of monitoring data sensed by a large number of collaborating sensors, is a challenging task because of many open technical issues, from typical WSN limitations (bandwidth, energy, delivery time, etc.) to the lack of widespread WSN data collection standards, needed for practical deployment in existing and upcoming IoT scenarios. In particular, effective collection is crucial for classes of smart city services that require a timely delivery of urgent data such as environmental monitoring, homeland security, and city surveillance. After surveying the existing WSN interoperability efforts for urban sensing, this paper proposes an original solution to integrate and opportunistically exploit MANET overlays, impromptu, and collaboratively formed over WSNs, to boost urban data harvesting in IoT. Overlays are used to dynamically differentiate and fasten the delivery of urgent sensed data over low-latency MANET paths by integrating with latest emergent standards/specifications for WSN data collection. The reported experimental results show the feasibility and effectiveness (e.g., limited coordination overhead) of the proposed solution.

Vehicular sensor networks are emerging as a new network paradigm of primary relevance, especially for proactively gathering monitoring information in urban environments. Vehicles typically have no strict constraints on processing power and storage capabilities. They can sense events (e.g., imaging from streets), process sensed data (e.g., recognizing license plates), and route messages to other vehicles (e.g., diffusing relevant notification to drivers or police agents). In this novel and challenging mobile environment, sensors can generate a sheer amount of data, and traditional sensor network approaches for data reporting become unfeasible. This article proposes MobEyes, an efficient lightweight support for proactive urban monitoring based on the primary idea of exploiting vehicle mobility to opportunistically diffuse summaries about sensed data. The reported experimental/analytic results show that MobEyes can harvest summaries and build a low-cost distributed index with reasonable completeness, good scalability, and limited overhead

The growing interest in the Internet of Things (IoT) has resulted in a number of wide-area deployments of IoT subnetworks, where multiple heterogeneous wireless communication solutions coexist: from multiple access technologies such as cellular, WiFi, ZigBee, and Bluetooth, to multi-hop ad-hoc and MANET routing protocols, they all must be effectively integrated to create a seamless communication platform. Managing these open, geographically distributed, and heterogeneous networking infrastructures, especially in dynamic environments, is a key technical challenge. In order to take full advantage of the many opportunities they provide, techniques to concurrently provision the different classes of IoT traffic across a common set of sensors and networking resources must be designed. In this paper, we will design a software-defined approach for the IoT environment to dynamically achieve differentiated quality levels to different IoT tasks in very heterogeneous wireless networking scenarios. For this, we extend the Multinetwork INformation Architecture (MINA), a reflective (self-observing and adapting via an embodied Observe-Analyze-Adapt loop) middleware with a layered IoT SDN controller. The developed IoT SDN controller originally i) incorporates and supports commands to differentiate flow scheduling over task-level, multi-hop, and heterogeneous ad-hoc paths and ii) exploits Network Calculus and Genetic Algorithms to optimize the usage of currently available IoT network opportunities. We have applied the extended MINA SDN prototype in the challenging IoT scenario of wide-scale integration of electric vehicles, electric charging sites, smart grid infrastructures, and a wide set of pilot users, as targeted by the Artemis Internet of Energy and Arrowhead projects. Preliminary simulation performance results indicate that our approach and the extended MINA system can support efficient exploitation of the IoT multinetwork capabilities.

This article investigates how and to what extent the power of collective although imprecise intelligence can be employed in smart cities. The main visionary goal is to automate the organization of spontaneous and impromptu collaborations of large groups of people participating in collective actions (i.e., participAct), such as in the notable case of urban crowdsensing. In a crowdsensing environment, people or their mobile devices act as both sensors that collect urban data and actuators that take actions in the city, possibly upon request. Managing the crowdsensing process is a challenging task spanning several socio-technical issues: from the characterization of the regions under control to the quantification of the sensing density needed to obtain a certain accuracy; from the evaluation of a good balance between sensing accuracy and resource usage (number of people involved, network bandwidth, battery usage, etc.) to the selection of good incentives for people to participAct (monetary, social, etc.). To tackle these problems, this article proposes a crowdsensing platform with three main original technical aspects: an innovative geo-social model to profile users along different variables, such as time, location, social interaction, service usage, and human activities; a matching algorithm to autonomously choose people to involve in participActions and to quantify the performance of their sensing; and a new Android-based platform to collect sensing data from smart phones, automatically or with user help, and to deliver sensing/actuation tasks to users.