Lightweight detection and classification for wireless sensor networks in realistic environments

Abstract

A wide variety of sensors have been incorporated into a spectrum of wireless sensor network (WSN) platforms, providing flexible sensing capability over a large number of low-power and inexpensive nodes. Traditional signal processing algorithms, however, often prove too complex for energy-and-cost-effective WSN nodes. This study explores how to design efficient sensing and classification algorithms that achieve reliable sensing performance on energy-andcost-effective hardware without special powerful nodes in a continuously changing physical environment. We present the detection and classification system in a cutting-edge surveillance sensor network, which classifies vehicles, persons, and persons carrying ferrous objects, and tracks these targets with a maximum error in velocity of 15%. Considering the demanding requirements and strict resource constraints, we design a hierarchical classification architecture that naturally distributes sensing and computation tasks at different levels of the system. Such a distribution allows multiple sensors to collaborate on a sensor node, and the detection and classification results to be continuously refined at different levels of the WSN. This design enables reliable detection and classification without involving high-complexity computation, reduces network traffic, and emphasizes resilience and adaptation to the realistic environment. We evaluate the system with performance data collected from outdoor experiments and field assessments. Based on the experience acquired and lessons learned when developing this system, we abstract common issues and introduce several guidelines which can direct future development of detection and classification solutions based on WSNs.

EnviroMic: Towards Cooperative Storage and Retrieval in Audio Sensor Networks

“…Sensor network applications have used acoustic sensors for different purposes, including localization [28], surveillance [12], communication [29], and geophysical monitoring [30]. Interestingly, none of these applications let users retrieve raw acoustic sensor samplings: they either use filtered samplings for application needs [12], or use acoustic signals for purposes other than recording [28].…”

On accurate and efficient statistical counting in sensor-based surveillance systems

“…For example, XSM motes [5] incorporate a band-pass filter to enhance the detection of acoustic emission, a digital potentiometer to detect a wide range of signals, and a polyethylene film to reduce the effect of sunlight. Besides hardware enhancements, advanced detection algorithms [6,7,8] have also been proposed to avoid misdetection with minimal energy consumption. VigilNet [7] utilizes a multi-level detection algorithm with in-situ adaptive thresholds to avoid both false positive and false negative detections in changing weather conditions.…”

“…Besides hardware enhancements, advanced detection algorithms [6,7,8] have also been proposed to avoid misdetection with minimal energy consumption. VigilNet [7] utilizes a multi-level detection algorithm with in-situ adaptive thresholds to avoid both false positive and false negative detections in changing weather conditions. Feng et al [8] propose a collaborative tracking algorithm with distributed Bayesian estimation to improve reliability based on current and previous estimation (beliefs) from sets of sensors.…”

“…Miss-detection is normally addressed by using reliable sensing hardware [5] and/or robust detection algorithms [6,7,8], while double-counting is a more challenging problem, because it involves duplicates in both communication and sensing. Several excellent projects have investigated how to avoid the double-counting problem in communication.…”

APL: Autonomous Passive Localization for Wireless Sensors Deployed in Road Networks

“…• Sensors have simple sensing devices for binary vehicle detection without any costly ranging or GPS devices [6,13]. Each detection consists of a sensor ID and timestamp, that is, (s i ,t j ) for i, j ∈ N.…”