An Efficient Intruder Detection using the Seismic Sensor

Abstract

This paper reports on a design of the footstep signal detection system using the seismic sensor. First, we analyzed the characteristics of seismic signal, seismic sensor, and the UGS(Unattended Ground Sensor) system with seismic sensors. In addition, we summarized the existing algorithms to detect footstep using the seismic sensors, and developed our low-power and high efficient footstep detection algorithm. In this paper, the sensor node operations are classified into three different steps and different resources and algorithms are applied to each step, not only to minimize the power consumption, but also to improve the performance.

Seismic Target Recognition Based on Parallel Recurrent Neural Network for Unattended Ground Sensor Systems

“…It is assumed that among thex 3 samples of the running signal, x 31 samples are recognized as no one, andx 32 samples are recognized as someone. According to the above assumption, the following equation is given as: (7) Since the test accuracy P 1 of the time domain feature network is obtained by the background signal and the running signal test, and the test accuracy P 2 of the frequency spectrum feature network is obtained by the background signal and the walking signal test, so the following equations are given as:…”

“…Nowadays UGS is widely used to monitor human activities, such as pedestrian movement and intruder detection in safe areas [2], [3]. The most commonly used sensor devices in UGS systems are micro-electromechanical-systems (MEMS) accelerometers [4], MEMS microphone array [5], quartz MEMS vibrating beam seismometer [6] and other seismic sensors [7]- [12]. The MEMS technology can integrate signal processing circuit and sensor element into a single chip, providing us with the control mechanism of sensor calibration and initial signal processing [4].…”