Seismic military sensors are required to be robust, reliable, compact, and easy to install and operate to be effective in the battlefield environment. Three types of sensor technologies were addressed that provide improved design and novel signal processing techniques: (a) a wavelength scanning, pulsed-laser-based demodulation system; (b) digital lock-in amplifier and field-programmable gate array (techniques) for weak signal detection and processing; and (c) improved seismic sensitivity based on carbon fiber optic composite cantilever and fiber-Bragg-grating (FBG).
A new scheme of laser-based optical demodulation was developed. At 100-Hz scanning speed, the wavelength was demodulated down to 1.1 pm. The signal was completely digital and clear, with excellent signal-to-noise ratio. At lower scanning speeds, sub-picometer wavelength resolution was achieved. The dynamic following range is about 10 nm, with a dynamic range of 120 dB.
The design of the seismic sensor is illustrated in the figure. The detection is implemented by the FBG dynamic strain sensor, which is attached on a spring-mass system. The acceleration of ground motion is transformed into strain variation on the FBG sensor through this mechanical design and, after the optical demodulation, generates the analog voltage output proportional to the strain changes. By adjusting the mechanical parameter of the spring-mass configuration, one can mechanically tune the natural response frequency of the system within a certain range in adapting to the different frequencies of seismic wave sources (signals of personnel and vehicles).
This sensor head has a compact size and a mass of only a few grams. The sensor can be embedded and hidden in the battlefield without any radio frequency emission or thermal signature to the environment. The optical fiber-based sensor itself is resistant to corrosion, high temperature, and fatigue, and is suitable for deployment in the harsh environment of the battlefield.
A damping mechanism is incorporated into the design with critical damping provided so that the mass-spring system will return quickly to its ready state after detecting a signal. This damping mechanism includes a Faraday induction loop and a permanent magnet that provide the damping. The small, induced current in the Faraday loop is properly sealed so that no electromagnetic signal will go in or out of the sensor head. By using the carbon fiber composite cantilever, the overall sensor performance is improved, which leads to a higher sensitivity, better linearity, and smaller weight on the sensor head.
The sensor is intrinsically waterproof and dustproof, but it is also very sensitive to temperature changes. It uses a pair of matched FBGs; one is the sensing element and the other is the demodulator. A thermostat and a temperature control make sure that the two FBGs are always kept at the same temperature.
Power consumption is another consideration for unattended system design. A solar cell is employed on the top cover of the sensor with a high-density battery inside the sensor. A digital control block permits cycling of wake-sleep configuration. Sleep time is set as long as possible based on average current consumption, and the ratio of wake-sleep is set depending on the required reliability of detection, alarm level, and the states of neighboring sensors.
One of the most profound challenges in the design of the seismic sensor is acquiring and separating weak seismic signals from strong background noise (for example, wind); another is distinguishing signals from different sources. Basic requirements for a practical unattended seismic sensor are that it has to be very sensitive to small vibrations, but also have a proper strategy to decrease the false alarm rate (FAR). To solve these problems, an original algorithm was designed.
The workings of this algorithm are illustrated by using the example of detecting the signals of a human walking. First, the system records the seismic response of a person walking, saving the basic digital model on the sensor. Second, the system compares the sensor’s subsequently acquired seismic signal with the basic digital model using a correlation operation. Third, the system estimates the degree of correlation. If the degree of correlation is higher than a preset level, it is counted as a signal impulse. Finally, the system sums the number of impulses in a given time interval. If this sum matches the rate of previous signals, the sensor initiates an alarm. Both improved signal detecting ability and decreasing false alarm rate are achieved with this algorithm.
This work was done by Joseph Dorleus of the U.S. Army Program Executive Office for Simulation, Training, and Instrumentation; and Yan Zhang, Jing Ning, Thomas Koscica, Hongbin Li, and H. L. Cui of the Stevens Institute of Technology. For more information, download the Technical Support Package (free white paper) at www.defensetechbriefs.com/tsp under the Physical Sciences category. ARL-0095
This Brief includes a Technical Support Package (TSP).
Fiber-Optic Seismic Sensor for Unattended Ground Sensing
(reference ARL-0095) is currently available for download from the TSP library.
Don't have an account? Sign up here.