This paper reports the results of a recently concluded R&D project, SCALS (Smart Cities Adaptive Lighting System), which aimed at the development of all hardware/software components of an adaptive urban smart lighting architecture allowing municipalities to manage and control public street lighting lamps. The system is capable to autonomously adjust street lamps’ brightness on the basis of the presence of vehicles (busses/trucks, cars, motorcycles and bikes) and/or pedestrians in specific areas or segments of the streets/roads of interest to reduce the energy consumption. The main contribution of this work is to design a low cost smart lighting system and, at same time, to define an IoT infrastructure where each lighting pole is an element of a network that can increase their amplitude. More generally, the proposed smart infrastructure can be viewed as the basis of a wider technological architecture aimed at offering value-added services for sustainable cities. The smart architecture combines various sub-systems (local controllers, motion sensors, video-cameras, weather sensors) and electronic devices, each of them in charge of performing specific operations: remote street segments lamp management, single street lamp brightness control, video processing for vehicles motion detection and classification, wireless and wired data exchanges, power consumptions analysis and traffic evaluation. Two pilot sites have been built up in the project where the smart architecture has been tested and validated in real scenarios. Experimental results show that energy savings of up to 80% are possible compared to a traditional street lamp system.

No abstract

No abstract

In real-world applications, audio surveillance is often performed by large models that can detect many types of anomalies. However, typical approaches are based on centralized solutions characterized by significant issues related to privacy and data transport costs. In addition, the large size of these models prevented a shift to contexts with limited resources, such as edge devices computing. In this work we propose conv-SPAD, a method for convolutional SPectral audio-based Anomaly Detection that takes advantage of common tools for spectral analysis and a simple autoencoder to learn the underlying condition of normality of real scenarios. Using audio data collected from real scenarios and artificially corrupted with anomalous sound events, we test the ability of the proposed model to learn normal conditions and detect anomalous events. It shows performances in line with larger models, often outperforming them. Moreover, the model’s small size makes it usable in contexts with limited resources, such as edge devices hardware.