The use of acoustic and seismic modalities in unattended ground sensor systems has been an active area of research. However, a thorough study of the advantages and complementary aspects of the two modalities has not heretofore been considered. This paper provides a detailed description of both acoustic and seismic methods for unattended ground sensors. We discuss acoustic and seismic phenomenology; source mechanisms, propagation paths, attenuation, and sensing. These phenomena significantly impact the detection, identification, and localization objectives of unattended ground sensors. As such, certain applications will benefit from exploiting acoustic ground sensing, seismic ground sensing, or in some cases, both. A methodology is presented for selecting the preferred modality (acoustic and/or seismic) for a particular applications. It is shown that the preferred modality significantly impacts the architecture of the ground sensor system. Important system architecture criteria such as transducer selection, data acquisition bandwidth, array geometry, and communications requirements are considered in this paper. Furthermore, the effects of seismo-acoustic and acousto-seismic interactions are addressed. It is shown that acoustic transducers may respond to seismic excitation, and seismic transducers often respond to acoustic excitation. In some instances, this dual-modality is of benefit. We also provide recommendations for minimizing seismic/acoustic cross talk for applications requiring independent seismic or acoustic modes.

Ground and air vehicles have distinctive acoustic signatures produced by their engines and/or propulsion mechanism. The structure of these signatures makes them amenable to classification by pattern recognition algorithms. There are substantial challenges in this process. Vehicle signatures are non-stationary by virtue of variations in engine RPM and maneuvers. Field sensors are also exposed to substantial amounts of noise and interference. We discuss the use of neural network techniques coupled with spatial tracking of the targets to carry out the target identification process with a high degree of accuracy. Generic classification is done with respect to the type of engine (number of cylinders) and specific classification is done for certain types of vehicles. This paper will discuss issues of neural network structure and training and ways to improve the reliability of the estimate through the integration of target tracking and classification algorithms.

This paper identifies and explores the technical requirements and issues associated with remotely monitoring continuous wave (CW) sources with seismic arrays. Potential approaches to this monitoring problem will be suggested and partially evaluated to expose the monitoring challenges which arise when realistic local geologies and cultural noise sources are considered. The selective directionality and the adaptive noise cancellation properties of arrays are required to observe weak signals while suppressing a colored background punctuated with an unknown distribution of point and sometimes distributive sources. The array is also required to characterize the emitters and propagation environment so as to properly focus on the CW sources of interest while suppressing the remaining emitters. The proper application of arrays requires an appreciation of the complexity of propagation in a non-homogeneous earth. The heterogeneity often limits the available spatial coherence and therefore the size of the array. This adversely impacts the array gain and the array's ability to carefully resolve various emitters. Arrays must also contend with multipath induced by the source and the heterogeneous earth. If the array is to focus on an emitter and realize an enhancement in the signal to noise ratio, methods must be sought to coherently add the desired signal components while suppressing interference which may be correlated with the desired signal. The impact of these and other issues on array design and processing are described and discussed.

Vibration sensors provide very important information for an unattended ground sensor network. For this specific application, a vibration sensor must be extremely robust and must have very high sensitivity and very low power consumption. The authors experimentally demonstrated two schemes of fiber optic interferometric sensors capable of sensing vibrations with amplitudes of a few tens of picometers in the frequency range below 1 kHz. The proposed schemes consist of commercially available components and make possible fabrication of a ruggedized, highly sensitive sensor with low power consumption.

A miniature absolute magnetometer is being developed by Raytheon Systems Company (RSC) for military applications under the Miniature Unattended Sensor Technology program. RSC's advanced magnetometer utilizes a laser to optically pump and spin polarize ³He in a small glass cell subjected to an applied field of two gauss. The spin polarized ³He nuclei process about the ambient field at a frequency proportional to the magnitude of the field. Through a process of relaxation, the spin polarized ³He becomes depolarized in a time called T₂. The precession generates a current in a pickup coil which is processed for a signal to detect, track and classify the passage of a mobile target. The magnetometer's unique characteristic of low noise at low frequencies makes the sensor well suited for detecting targets at long range.

The detection of nitroaromatic molecules by porous silicon films has been explored using direct and indirect detection methods. In the direct method, detection is achieved by monitoring the photoluminescence of a nanocrystalline porous Si films upon exposure to the analyte of interest. Photoluminescence is quenched upon adsorption of the nitroaromatic, presumably via an electron transfer mechanism. For nitrobenzene a detection limit of 350 ppm (after an exposure time of < 2 minutes) was observed. For 2,4-dinitrotoluene, a much lower detection limit of 250 ppb (after an exposure time of < 6 minutes) was obtained. Both the detection limit and the response time of the material can be lowered by the use of a catalyst (PtO₂ at 250 degree(s)C) in the carrier gas line upstream of the porous silicon detector. The enhanced sensitivity comes from catalytic oxidation of the nitroaromatic to NO₂, which irreversibly oxidizes the surface of the porous Si, providing an integrating function. The demonstrated limit for NO₂ detection is 70 ppb. A complementary detection technique involving measurement of spectral shifts of a porous Si film Fabry-Perot interferometer upon oxidation will also be presented.

The ability to characterize suspect facilities for intelligence or counterforce purposes will rely heavily on the ability to identify chemical effluents from such facilities. Sandia is developing a complete micro-scale chemical analysis system named (mu) ChemLab^TM. This system will be extremely small and low power because of the utilization of integrated circuit fabrication techniques. The use of monolithic integration of such components as chemical preconcentration, separation, and detection, along with the eventual integration of micromachined pumps and valves, will not only lead to a dramatic size reduction, but also lead to the dramatic cost reduction that has been realized with monolithic integration of electronics. It will become practical to deploy large numbers of unattended chemical analysis systems for sensing low concentration effluents at high priority targets. (mu) ChemLab^TM uses an array of serial and parallel separations channels (columns), each of which separates compounds on the basis of different `orthogonal' chemical properties, followed by highly sensitive detection techniques: laser-induced fluorescence in the liquid phase and arrays of acoustic wave devices in the gas phase. This array of separations will create a characteristic, highly specific signature for a acoustic wave devices in the gas phase. This array of separations will create a characteristic, highly specific signature for a compound. Identification of target species based on the combined results of multiple separations will have an extremely low false alarm rate because each separation is statistically independent. Because the separations will be run simultaneously in microchannels, analyses times are on the order of a few minutes. The necessary sample handling and detection systems will be implemented using microfabricated electronic, optical and fluidic components.

A unique corrugated-thin-metal-film structure suitable for integration with semiconductor photodetectors, is investigated. For fluorescent spectral monitoring, this structure provides a number of important detection features including fluorescence enhancement for molecules located near the surface of the sensor, excitation wavelength shielding and narrow bandpass filtering of detected wavelengths. Surface dielectrics with chemical or biological-specific properties may also be incorporated into the sensor for additional selectivity.

There exists a need to develop novel, advanced, unattended magnetic and chemical micro-sensor systems for successful detection, localization, classification and tracking of ground time critical targets of interest. Consistent with the underlying long-term objectives of the development of unattended ground sensors program we have investigated the use of a new planted ground sensor platform based on Micro- Electro-Mechanical Systems that can offer magnetic, chemical and possibly acoustic detection. The envisioned micro-system will be low-power and low-cost and will be built around a single type of microstructure element integration a monolithic optical system and electronics package. This micro sensor can also incorporate burst telemetry to transmit the information, a renewable power source and will be capable of operating under field conditions, with sufficient sensitivity to permit high detection rates, and with sufficient chemical selectivity to prevent high false alarm rates. Preliminary studies, initial designs, and key predicted performance parameters will be presented. Possible applications of such as system include sensitive perimeter monitoring such as minefields and military/nuclear bases, vehicle detection, and aircraft navigation systems, and drug enforcement operations. The results of the present work demonstrate that the microcalorimetric spectroscopy technique can be applied to detect and identify chemicals in the ppm level and the studied microcantilever-based magnetometer can provide sensitivities in the order of 1(mu) T.

This paper describes the UL3 camera system based on the advanced 160 X 128 uncooled micro-bolometer FPA. The UL3 camera takes advantage of the patented bias equalization FPA performance techniques to produce the world's smallest IR camera. The UL3 camera weights less than 2.3 ounces, uses less than 600 mW of power, and has overall dimensions of 3 cm X 3 cm X 6 cm. The architectures feature an approach that integrates the required system functions on ASICs and FPGAs rather than including discrete components and microprocessors.

Low power and low cost are primary requirements for an imaging infrared camera used in unattended ground sensor arrays. In this paper, an amorphous silicon (a-Si) microbolometer-based uncooled infrared camera technology offering a low cost, low power solution to infrared surveillance for UGS applications is presented. A 15 X 31 micro infrared camera (MIRC) has been demonstrated which exhibits an f/1 noise equivalent temperature difference sensitivity approximately 67 mK. This sensitivity has been achieved without the use of a thermoelectric cooler for array temperature stabilization thereby significantly reducing the power requirements. The chopperless camera is capable of operating from snapshot mode (1 Hz) to video frame rate (30 Hz). Power consumption of 0.4 W without display, and 0.75 W with display, respectively, has been demonstrated at 30 Hz operation. The 15 X 31 camera demonstrated exhibits a 35 mm camera form factor employing a low cost f/1 singlet optic and LED display, as well as low cost vacuum packaging. A larger 120 X 160 version of the MIRC is also in development and will be discussed. The 120 X 160 MIRC exhibits a substantially smaller form factor and incorporates all the low cost, low power features demonstrated in the 15 X 31 MIRC prototype. In this paper, a-Si microbolometer technology for the MIRC will be presented. Also, the key features and performance parameters of the MIRC are presented.

We describe tomographic integration of visible and IR projections to form 3D models of object spaces. 3D modeling is a computationally efficient means of multiple sensor data reduction and a robust means of scene analysis. We consider how imaging system design might be revised to obtain 3D data and present experimental data from a compact interferometric coherence imaging system. We experimentally and theoretically analyze 3D reconstructions of macroscopic objects, focusing in particular on the reconstruction of an automobile using a conventional camera.

A system has been developed for delivering and attaching a sensor payload to a target using a standard 40-mm grenade launcher. The projectile flight characteristics are similar to existing 40-mm rounds, with a useful range of up to 300 m. The projectile incorporates an attachment mechanism, a shock mitigation system, a power source, and a transmitter that allows sensor data to be transmitted to a receiver at up to 1/4 mile range. Impact g-loads have been limited to less than 10,000 g's, enabling sensor payloads to be assembled using Commercial Off-The-Shelf components. The GLIMPS projectile is intended to be a general purpose delivery system for a variety of sensor payloads under the Unattended Ground Sensors program, with Phase I proof-of- concept being demonstrated using a low-power CMOS camera.

The Organic Sensor Array program is investigating the application of small imaging sensors to support Small Unit Operation teams and Unattended Ground Sensor applications. Experiments are underway for emplaced and sense-on-the-fly modes of operation. These prototypes are of interest to the UGS community for adding imaging surveillance in a micro- package. Imaging provides for confirmation of events and identification of threats detected by other UGS sensors. Performance to data has shown progress in attachment, shock mitigation, flight stability and video processing.

This report covers work under phase one of the Micro Unattended Mobility System project investigating the addition of a mobile sensor components to existing and future ground penetrator delivered unattended sensor systems. A typical unattended sensor strategy consists of air-dropping sensor packages into a target terrain for remote observation and intelligence gathering. Existing and planned unattended systems have no control over their location after the drop is complete. We propose to augment the capability of these sensing packages by giving them a degree of local mobility. From an assumed operational scenario, vehicle design specifications are identified that would be required for mission success. Three basic mobility concepts are presented and evaluated for their strengths and weaknesses in the proposed mission. The mobility concepts are grouped into wheeled, jumping, and crawling systems. Of the three mobility concepts discussed, the system that shows the most promise is presented in a more detailed design. This design consists of two side by side wheels which drag a reaction tail behind them. The control electronics, batteries, and drive motors are housed in a central body connected to the tail and two sensor payloads can be placed in the wheel hubs. This design is proposed for further development and testing in the second phase of this project.

Measurement and signal intelligence of the battlespace has created new requirements in information management, communication and interoperability as they effect surveillance and situational awareness. In many situations, stand-off remote-sensing and hazard-interdiction techniques over realistic operational areas are often impractical and difficult to characterize. An alternative approach is to implement adaptive remote-sensing techniques with swarms of mobile agents employing collective behavior for optimization of mapping signatures and positional orientation (registration). We have expanded intelligent control theory using physics-based collective behavior models and genetic algorithms to produce a uniquely powerful implementation of distributed ground-based measurement incorporating both local collective behavior, and niter-operative global optimization for sensor fusion and mission oversight. By using a layered hierarchical control architecture to orchestrate adaptive reconfiguration of semi-autonomous robotic agents, we can improve overall robustness and functionality in dynamic tactical environments without information bottlenecking.

We begin by considering current shortfalls with conventional surveillance systems and discuss the potential advantages of distributed, collaborative surveillance systems. Distributed surveillance systems offer the capability to monitor activity from multiple locations over time thereby increasing the likelihood of obtaining discriminating data necessary for interpretation of the activity. Yet the multiplicity of sensors magnifies the volumes of data that must be processed. We present our vision of a system which generates timely interpretations of activities in the scene automatically through the use of mechanisms for collaboration among sensing systems and efficient perception methods which complement the sensing paradigm. Then we review our recent efforts toward achieving this goal and present initial results.

Operational guidance on reliability as a function of site conditions is essential to unattended ground sensors (UGS) effectiveness. For physical security/force protection sensor systems, the Weather Vulnerability Assessment Tool (WVAT) will fill this need. WVAT is a computer application that will warn of situations in which the likelihood of detecting an intruder falls below a specified probability of detection (P_d). Both P_d and the likelihood of nuisance alarms are predicted for diurnal and seasonal variability in weather, state of the ground, and ground cover. A follow-on computer applications, the Force Protection Sensor Selector, will guide the selection of sensor systems for deployment as a site, based on its prediction of weather-induced variation in site conditions.

Tactical sensing on the battlefield involves real-time information processing of acoustic, seismic, electromagnetic, and environmental sensor data to obtain and exploit an automated situational awareness. The value-added of tactical sensing is to give the war-fighter real-time situational information without requiring human interpretation of the underlying scientific data. For example, acoustic, seismic, and magnetic signatures of a ground vehicle can be used in a pattern recognition algorithm to identify a tank, truck, or TEL, but all the war-fighter wants to know is how many of each vehicle type are present and which way are they going. However, the confidences of this automated information processing are dependent on environmental conditions and background interference. A key feature of tactical ground sensing is the ability to integrate objective statistical confidences into the process to intelligently suppress false alarms, thus allowing the war-fighter to concentrate on war fighting. This paper presents how the situation confidence metric is generated starting with the sensor SNR going all the way through to the target classification and track confidences. This technique also allows modeling of ground sensor performance in hypothetical environments such as bad weather.

In this paper, we quantify how communication increases the effective range of detection of unattended ground sensors. Statistical analysis is used to evaluate the probability of detection for multiple sensors using one, two, and infinite levels of cooperation. Levels of cooperation are defined as the levels of communication between sensors. One level of cooperation means that one sensor passes its state information to several other sensors within a limited communication range, but this information is not passed beyond this range. Two levels of cooperation means that the state information received by this first set of sensors is relayed to another set of sensors within their communication range. Infinite levels of cooperation means that the state information is further percolated out to all sensors within a communicating group. With large numbers of sensors, every sensor will have state information about every other sensor regardless of communication range. With smaller numbers of sensors, isolated groups may form, thus lowering the probability of information transfer.

Unattended ground sensors (UGS) provide surveillance, intelligence, and monitoring of areas not suitable for continuous human presence. Current state-of-the-art UGS are larger and heavier than desired and usually are limited to one or two types of sensors. Northrop Grumman is developing the next generation UGS which is called the Miniaturized Unattended Ground Sensor (Mini-UGS). The Mini-UGS will employ a modular concept which allows the UGS to be reconfigured as mission requirements change. Multiple types of sensors and/or radios can be interchangeable on the Mini- UGS. In order to make the Mini-UGS a reality, several critical technologies such as low power processor architectures, rugged communication bus, and packaging hardening techniques are under development. This paper presents the Mini-UGS concept and the critical technologies of which it is composed.

Advances in CMOS IC and micro electrical-mechanical systems (MEMS) technologies are enabling construction of low-cost building blocks each of which incorporates sensing, signal processing, and wireless communications. Collections of these integrated microsensor nodes may be formed into sensor networks in a wide variety of ways, with characteristics that depend on the specific application--the total number of nodes, the spatial density, the geometric configuration (e.g., linear vs. areal), topographic aspects (e.g., smooth vs. rough terrain), and proximity and proportion of user/sink points. The power of these distributed sensor networks will be unleashed by means of their ability to self-organize, i.e., to bootstrap and dynamically maintain organizational structure befitting the purpose and situation that is presented, without the need for human assistance. A prototype sensor system and networking protocols are being developed under the DARPA/TTO AWAIRS Program and are described.

The capabilities of unattended ground sensors (UGSs) have steadily improved and have been shown to be of value in various military missions. Today's UGS are multi-functional, integrated sensor platforms that can detect and locate a wide variety of ground-based and airborne targets. The rather large size (> 1 cubic foot) and relatively expensive cost of these integrated platforms are two main drawbacks for remote surveillance applications that support rapidly deployable, small unit operations. As an alternative, remote surveillance may be possible with smaller, less costly sensors that work cooperatively together as a network. The objective of this study was to develop algorithms that can optimally organized and adaptively control a network of UGSs in order to achieve a surveillance mission. In the present study, the sensor network, a random distribution of acoustic sensors over a surveillance area, is tasked to detect and track any targets entering into the surveillance area. In addition, the sensor network is required to maximize its tracking accuracy and minimize its power utilization.

Adding remote deployment capability to unattended surveillance systems provides the capability to unobtrusively monitor activities in hostile or neutral areas where in-situ placement is not practical. Targeted surveillance activities include tracking force movements, monitoring keep-out areas, and detecting the presence of assets hidden underground or in buildings. Key technologies to achieve this capability include localization of deployed nodes, field optimization and control techniques, multi- sensor fusion, and low cost miniaturized sensors.

Distributed microsensor networks, built from collections of nodes each having the ability to sensor their environment, process the raw sensor data in cooperation with other neighboring nodes into information and then communicate that information to end users. These systems are designed to be self-organizing in the sense of establishing and maintaining their own network without the need for specialistic operators. In most envisioned applications, wireless communications are the most practical means of interconnection, eliminating the internode cabling. Long periods of autonomous operations in remote environments will need battery or other renewable energy sources. In order to prolong battery life, all node hardware and software functions need to be designed to consume minimal power. In general, a node will expend energy on local processing of sensor data to produce compressed information in order to reduce communications. These network systems are intended to support large numbers of such nodes to cover large geographic areas.