A distributed sensor management for large-scale IoT indoor acoustic surveillance

摘要

The recent world events have underscored the need for large area surveillance systems. Such systems require effective sensing and collaborative decision-making to operate in highly dynamic environments with demanding time constraints. The Pervasive Internet of Things (IoT) is a novel paradigm that enables detailed characterization of the real physical applications. To this end, a pervasive IoT surveillance applications can offer an effective framework to collect situation-aware knowledge that is vital for planning effective security measures. Nevertheless, most state-of-the-art focus only on visual abnormal event recognition using centralized systems, thus, ignoring the need for distributed operation to enable large-scale IoT surveillance systems. This paper presents a novel Sensor Management (SM) framework for pervasive IoT acoustic surveillance, IntelliSurv, that automatically detects and localizes abnormal acoustic events in a distributed collaborative manner. The proposed framework coordinates the sensing resources using a novel team-theoretic SM, based on the Belief–Desire–Intention (BDI) model, for autonomous decision-making and resource allocation. The proposed abnormal event recognition module, using Support Vector Machines (SVM) and Linear Discriminate Analysis (LDA) classifiers, relies on audio information to recognize human screams or high-stress speech signals. The simulation scenario in this work is the surveillance of the Waterloo International Airport implemented using Jadex platform and Speech Under Simulated and Actual Stress (SUSAS) database. The simulation results show the merits of the proposed IntelliSurv framework, compared to the popular centralized systems, over varying network size, number of threats, Signal-to-Noise Ratios (SNR), tracking quality, and energy consumption.