6th Future SecuritySecurity reSearch conFerenceBerlin, SeptemBer 5th 7th, 2011

proceedingSJoachim Ender, Jens Fiege (Eds.)

F r a u n h o F E r g r o u p F o r d E F E n s E a n d s E c u r i t y

W

olfg

ang

Kla

tt

editorial noteS

ISBN 978-3-8396-0295-9

9 7 8 3 8 3 9 6 0 2 9 5 9

Editors

Joachim Ender Jens Fiege

contact

Fraunhofer Institute for High Frequency Physics and Radar Techniques FHRNeuenahrer Str. 20 53343 Wachtberg, Germany www.fhr.fraunhofer.de

Phone + 49 228 9435-227 Fax + 49 228 9435-627 E-Mail info@fhr.fraunhofer.de

conference host

Fraunhofer Group for Defense and Securitywww.vvs.fraunhofer.de

Bibliographic information published by Die Deutsche Bibliothek

Die Deutsche Bibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliografic data is available in the Internet at http://dnb.d-nb.de.

ISBN 978-3-8396-0295-9

All rights reserved; no part of this publication may be translated, reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the written permission of the publisher.Many of the designations used by manufacturers and sellers to distinguish their products are claimed as trademarks. The quotation of those designations in whatever way does not imply the conclusion that the use of those designations is legal without the consent of the owner of the trademark.

by FRAUNHOFER VERLAG, 2011Fraunhofer Information-Centre for Regional Planning and Building Construction IRBP.O. Box 80 04 69, D-70504 StuttgartNobelstrasse 12, D-70569 StuttgartPhone +49 711 970-2500Fax +49 711 970-2508E-Mail verlag@fraunhofer.deURL http://verlag.fraunhofer.de

proceedingSJoachim Ender, Jens Fiege (Eds.)

FraunhoFer VVS

6th Future SecuritySecurity reSearch conFerenceBerlin, SeptemBer 5th 7th, 2011

taBle oF contentS

proceedings

a.1 sensor technology for security

Scanning for Hazardous Objects on the Seafloor State of the Art Technologies ........................................................................1 Detection of High Power Microwaves .........................................................................................................................................8 Clinotrons High Power Sources for Terahertz Sensors .............................................................................................................14 Scanning Polarimetric Imaging Radiometer: Microwave Imaging System and Image Merging with IR and Optical Data ..............18 Standoff Detection of Suicide Bombers in Mass Transit Environment .....................................................................................24

B.1 crisis Management i

Automated Planning in Evolving Contexts: an Emergency Planning Model with Traffic Prediction and Control ...........................28 Highly Efficient Event and Action Processing for Emergency Management in Large Infrastructures ............................................38 Coordinating Ambulance Operations ........................................................................................................................................47 PROSIMOS A Tool for Identifying Business Cases in the Implementation of a Priority Communications Systems for First Responders in Public Mobile Networks .........................................................................................................................51 An Integrated and Integrating Airport Security Management Concept ......................................................................................60

a.2 supply chain security (invited)

Supply Chain Integrity Services Based on Hierarchical Sensor Networks .....................................................................................66 ProAuthent Integrated Protection Against Counterfeiting in Mechanical Engineering Through Marking and Athenticating Critical Components ...........................................................................................................................................71 Developing an Understanding of Supply Chain Security Management .......................................................................................75 100% Container Scanning: Impact on Efficiency and Costs of Container Terminal Operation ....................................................79

B.2 crisis Management ii

SECURITY2People Features of and Experience With the First Demonstrator of an Integrated Disaster Management System ................................................................................................................................................................83 Crowd Management Simulation Crowd Management in Large Infrastructures .......................................................................87 FP7 Project ACRIMAS Aftermath Crisis-Management System-of-Systems Demonstration ........................................................93 Process Structures in Crises Management ...............................................................................................................................101

posters Universal Detector of Concealed Hazardous Materials .............................................................................................................106 Laser Ion Mobility Spectrometer Technology and Security Applications ....................................................................................112 Fluorescent Biosensors for Standoff-Detection of Gamma-Radiation .......................................................................................117 Development of a Fully Automated Centrifugal Lab-on-a-Chip System for Rapid Field Testing of Biological Threats .................121 Detection Technologies Common Concepts in Security and Safety .......................................................................................125 Variable Irradiation Geometry With a New X-Ray Backscatter Camera for Security Applications ...............................................128 Surface Sensitive Detection of Trace Explosives With UV Photofragmentation .........................................................................132 A Security and Surveillance Solution for Scenarios With Time-Critical Response Time ..............................................................140 Electromagnetic Protection of IT-Networks for Transportation-Infrastructure (EMSIN) ...............................................................146

posters

Managing Security Tasks With Modular and Mobile Sensor Data Processing Networks An Integral Approach .......................152 Towards Smart Infrastructures For Modern Surveillance Networks ...........................................................................................158 Application of Special Purpose Blast Sets For Personal Rescue in a Hazardous Environment .....................................................164 Elimination of a Tanker Fire Through Shock Wave Interference ................................................................................................168 Acoustic-Generator Based on a Small Rocket-Burner With Intermittent Combustion to Dissolve Violent Demonstrations .........178 FP7 Project ETCETERA - Evaluation of Critical and Emerging Technologies for the Elaboration of a Security Research Agenda ..181 Presentation of TALOS, a Project of a Mobile, Scalable and Autonomous System for Protecting European Borders ..................186 Risk Treatment Measures for Managing Cargo Theft in Road Transportation ...........................................................................195 Risk Analysis for a German Harbour within the Project ECSIT ..................................................................................................201 Integrated Open-Source Software for Modeling the Effects of Bio- or Agroterroristic Attacks on The Food Chain ....................208 Laser-Based Ranging and Tracking of Space Debris..................................................................................................................209 Concept for the Integration of Predictive Microbiology Tools and Models in the Efforts to Secure the Food Supply Chain in Case of Bioterroristic Attacks ....................................................................................................................................215 Scenario-Oriented Assessment of Hazardous Biological Agents ...............................................................................................216 Positioning and Tracking of Deployment Forces Combining an Autonomous Multi-Sensor System with Video Content Analysis ...........................................................................................................................................................220 Sensors Data Fusion and Management in a New Security System on Airports ..........................................................................225 Data Protection and Security Awareness in Complex Information Systems ...............................................................................230 Efficient and Secure Data Transfer Using Jpeg Image Based Steganography ............................................................................240 Impact of Jamming on a Security-Enabled Anonymous MANET Protocol (SEAMAN) ................................................................246 Enhancing Information Security with Universal Core Approach ...............................................................................................251 A New System for Mobile Phone Localization for Search and Rescue Applications ...................................................................257 Multistatic 96 GHz Rotating W Band Radar for Passenger Inspection on Airports ....................................................................261 A Multichannel Scanning Receiver System for Surveillance Applications ..................................................................................265 How to Model and Simulate Multi-Modal Alerting of Population: The Alert4All Approach ......................................................269 VALUESEC - Mastering the Value Function of Security Measures .............................................................................................277 A Historical Analysis on the Nature of Criminal and Terrorist Threats Against Civil Aviation ......................................................282 Esfo The Information System on European Security Research ................................................................................................287

a.3 detection of hazardous Material

Novel Sensor Platform for Multiplexed Trace Detection of Hazardous Substances ....................................................................289 Change Detection on Millimeter-Wave SAR Images for C-IED Applications ..............................................................................293 Detection and Identification of Illicit and Hazardous Substances with Proton-Transfer-Reaction Mass Spectrometry (PTR-MS) ..298 Characterisation of Critical Material Based on Phase and Amplitude Information of High Frequency Measurements ................304

B.3 Video surveillance (invited)

Visual Search in Large Surveillance Archives ............................................................................................................................310 Towards People Re-Identification in Multi-Camera Surveillance Systems ..................................................................................315 Multi-Spectral and Hyperspectral IR-Sensors for Improved Surveillance Applications ................................................................321 Automatic Maritime Video Surveillance with Autonomous Platforms ......................................................................................326

taBle oF contentS

a.4 Maritime security

Introduction to Anti-Piracy The EU Operation Atalanta .........................................................................................................332 Development of Indicators to Evaluate a Vessels Vulnerability to Pirate Attacks and Packages of Appropriate Technological Protection Systems ............................................................................................................................................339 Polarimetric Detection of Small Maritime Targets for Maritime Border Control .........................................................................346 Handling Security Relevant Information in the Maritime Domain with the Security Modeling Technique ..................................350 New Challenges for Maritime Safety and Security Training Presentation of a Specific Safety & Security Trainer (SST7) ..............356

B.4 social dimension of security

Psychosocial Support for Civil Protection Forces Coping with CBRN An EU-Project ................................................................366 Leaking in the Name of Justice ...............................................................................................................................................370 Enhancing the Acceptance of Technology for Civil Security and Surveillance by Using Privacy Enhancing Technology ...............372 Customer Security Environment: Understanding Customers Views on Security .......................................................................380 Towards Information Services for Disaster Relief Based on Mobile Social Networking...............................................................386

a.5 radar sensors for security awareness

Security in Space Space Situational Awareness via Radar Observation ...................................................................................395 Ground Moving Target Indication And Ship Surveillance With The German Terra SAR-X/TanDEM-X Radar Satellite Constellation ...................................................................................................................................................401 SecurityRelated Change Detection with TerraSAR-X Radar Satellite Data ...............................................................................408 Pulse Radar Technology for Detection of Trapped and Buried Victims Electronic Devices ..........................................................412 An Integrated Radar-Optronic Sensor Architecture and OperationalExperiences ....................................................................416

B.5 anomaly detection and risk analysis

Towards Proactive Security Surveillance by Combining Technology and Human Factors ...........................................................421 Searching for Abnormalities Instead of Suspects .....................................................................................................................426 Applied Text Mining for Military Intelligence Necessities ..........................................................................................................431 Topic-Oriented Analysis of Data Streams .................................................................................................................................438 Video Analysis for Situation and Threat Recognition ...............................................................................................................443

a.6 terahertz security applikations (invited)

Security Check of the Future ...................................................................................................................................................447 Progress in Device Technology Creates Potential for Active Real-Time THz Security Scanners ...................................................453 Terahertz Sensor Systems for Field Applications .......................................................................................................................457 QPASS Quick Personnel Automatic Safe Screening for Security Enhancement of Passengers .................................................462 Millimeter Wave Radar Sensor for Protection of Outdoor Areas ...............................................................................................468

B.6 critical infrastucture

Geometrical Design Criteria for Analyzing the Vulnerability of Urban Area Construction to Blast Effects ..................................472 Security Impact Simulation for Critical Infrastructure of Freight Villages Using Software-Agents ..............................................479 Safety and Protection of Built Infrastructure to Resist Integral Threats (SPIRIT) .........................................................................485 Risk Evaluation for Critical Built Infrastructure Asset Classification and Evaluation ...................................................................490 Servitization in Security Business .............................................................................................................................................498

a.7 response to cBrnE threats

Research Against CBRN-E Terrorism: A Real Opportunity for Materials Science ........................................................................503 DECOTESSC1: Results of an EU FP7 Demonstration Project Phase 1 CBRNE System-of-Systems Analysis ..................................513 EXAKT Joint BMBF Research Project: Near Real-Time Trace Analysis of Airborne Chemical Warfare Agents and Explosives using a TD-GC-TOF-MS ..........................................................................................................................................519

B.7 Border security (invited)

Integrated Border Management - Remarks on a Border Control Roadmap ..............................................................................525 Coastal Surveillance Radars Developed in TUBITAK BILGEMUEKAE ..........................................................................................530 Enhancing Nuclear Security at Ukrainian Border Stations to Prevent Illicit Trafficking ................................................................535

a.8 Food chain and transport security

Risk Assessment, Epidemiology, Detection of Biological Agents to Secure the Feed and Food Chain........................................539 Network of German Authorities in the Context of Bioterrorism in the Food Chain ..................................................................540 Securing the Feed and Food Supply Chain in the Event of Biologicaland Agro-Terrorism (BAT) Incidents The German SiLeBAT Project ...................................................................................................................................................542 Improving the Security of Critical Transport Infrastructures New Methods and Results ..........................................................545 Improving Security in Intermodal Transports ............................................................................................................................555

B.8 cyber defense and information security

Botnets: Detection, Measurement and Defense ......................................................................................................................562 Realising a Trust Worthy Sensor Node with the Idea of Virtualisation .......................................................................................568 WSNLab A Security Testbed for WSNs ..................................................................................................................................575 Interoperability of Information Systems for Public Urban Transport Security: The SECUR-ED Approach.....................................580 Security and Backup-System at the IT Center of the Technical University of Applied Science Wildau Including Autonomous Satellite Faculty and Degree Programme IT Systems ................................................................................................................586

a.9 Multiple sensor checkpoint control (invited)

Detection, Classification and Localization of Hazardous Substances in Public Facilities .............................................................590 Multisensory Acquisition for Situation Awareness in Riot Control Scenarios .............................................................................594 The Need for High-Performance Detectors in Security Applications: Results from a Test Bed for the Detection of Vapours Emitted from Moving Sources and the Results from Outgassing Experiments of Packaged TATP ..............................................597 Detection of Explosives Scenarios, Sensors and Realistic Concentrations ...............................................................................604 Multi-Sensor Awareness for Protection and Security ................................................................................................................608

B.9 Surveillance and Identification of People (invited)

GPS/EGNOS Based Surveillance and Guidance in an Airport Environment ................................................................................612 Realtime Event Detection and Prediction on Position Data Streams ..........................................................................................618 Security Systems With Seamless Authentication Based on Smart Phones and Surveillance Cameras .........................................622 A Step Forward to Automated Latent Fingerprint Segmentation .............................................................................................627

Scanning for Hazardous Objects on the Seafloor State of the Art Technologies Wolfgang Jans1), Holger Schmaljohann1), Florian Langner1, 2), Christian Knauer2), and Wolfgang Middelmann 3) 1) Bw Technical Center for Ships and Naval Weapons, Naval Technology and Research (WTD71), Research De-partment for Underwater Acoustics and Marine Geophysics (FWG), Klausdorfer Weg 2-24, 24148 Kiel, Ger-many, e-mail: WolfgangJans@bwb.org 2) Universitt Bayreuth, Institut fr Informatik, AG Algorithmen und Datenstrukturen, Universittsstrae 30, 95447 Bayreuth, Germany, e-mail: christian.knauer@uni-bayreuth.de 3) Fraunhofer Institute of Optronics, System Technologies and Image Exploitation (IOSB), Gutleuthausstrae 1,76275 Ettlingen, Germany, e-mail: Wolfgang.Middelmann@iosb.fraunhofer.de

Abstract

A broad variety of objects can be found on the seafloor. This starts for example with sea mines, a cheap and widely used weapon, followed by wrecks, waste, or dumped ammunition and ends with pipelines or underwater archeological finds. Sea mines pose a very effective threat to Navies and the free maritime trade. Besides conventional units Navies worldwide consider increasingly unmanned underwater vehicles (UUVs) with new technological solutions for mine counter measures. This includes in particular high-resolution sonar techniques such as synthetic aperture sonar (SAS) and computerized image analysis. Simultaneously, these new technologies can be adapted for civil applications. The Synthetic Aperture Sonar (SAS) is a new, innovative development in the field of acoustic imaging of the sea-bed, which is related to Synthetic Aperture Radar (SAR). By coherent addition of data from a series of consecu-tive pings a significantly longer antenna is synthesized in the direction of travel. Hence, the lateral resolution is often improved by an order of magnitude or even more for SAS Systems compared to conventional side scan sonars. This paper will discuss briefly several main differences between SAR and SAS and the consequences. Due to the much longer time for traveling one synthetic aperture, motion compensation is perhaps one of these differences. An other main difference is the stronger influence of present environmental conditions. E.g., the image quality can decline in wide areas due to the influence of multipath propagation mainly caused by reflections at the sea surface or sea floor. Then we will present our ongoing research in object detection and classification based on SAS image data. This aims at a completely autonomous object detection and mine classification approach ("Automatic Target Recogni-tion ATR)" for unmanned underwater vehicles. Our software is divided in into a couple of processing steps start-ing with pre-processing, screening for regions of interest, reduction of false positives, object classification up to and including fusion of detection and classification results for different algorithms. Examples for these different processing steps based on real SAS data will be presented and discussed.

1 Introduction

Sea mines are a cheap and widely used weapon. They pose a very effective threat to Navies and the free ma-ritime trade. Therefore, detecting and / or eliminating mines at sea is one of the most important tasks of any navy in order to keep sea routes open. Every Side Scan Sonar (SSS) allows for imaging large areas of the sea floor with a relatively high resolution in relatively short time. Therefore, SSS systems are

very interesting for military purposes and have been constantly improved since the 1950s. First used to e.g. find H-Bombs lost at sea or a lost Russian submarine [1], today modern synthetic aperture SSS systems are used to detect and classify small underwater objects including underwater threads like sea mines or impro-vised explosive devices (UW-IEDs). SSS systems are particularly well suited for Unmanned Underwater Vehicles (UUVs). Apart from military purposes, SSS imagery can be

Session A.1Future Security 2011 Berlin, September 5-7, 20111

used for various additional applications. E.g., among other sensors, a low frequency 120/410kHz EdgeTech 2200-MP SSS was used to locate the debris field and bodies of flight AF447, which disappeared on May 31st 2009 over the Mid-Atlantic at 02:14 UTC, in approx. 3900m depth after a week of a search and re-covery missions in April 2011. This SSS sonar on a REMUS 6000 AUV from IFM GEOMAR in Kiel operated 75m above the seafloor and mapped 750m to both sides of the vehicle track during surveys [2]. Apart from the ABYSS AUV of IFM GEOMAR two additional REMUS 6000 AUVs from the Waitt Insi-tute for Discovery / Wood Hole Oceanograthic Institu-tion were deployed during these missions. But the main field of SSS imagery produced by UUVs is the shallow water environment - although the hard-ware has to be adopted compared to deep sea equip-ment. As an example, a detailed SSS image of a Dorn-ier 17 WW II bomber wreck at a depth of ca. 16 m can be found under [3]. This bomber was damaged during the Battle of Britain in August 1940 and at-tempted an emergency landing on the Goodwin Sands, a sand bank in the English Channel. The aircraft ground looped during landing and sank inverted. Ad-jacent to the wreck a small debris field was found, comprising e.g. flaps and bomb bay doors, torn off during landing. The wreck lay proud of sand than dis-covered in a tidal area with low visibility (max. 5 m). In addition to entire structures and air planes current conventional SSS imagery can also be used to search for small objects and bodies. As an example the search and recovery result, looking for a drowning victim, can be found under [4]. Possible oil drums within a debris field in a harbor (see [5]) are a second example, which shows the potential to protect the environment by using SSS systems. Using Synthetic Aperture Sonar (SAS) processing further improves this capability of SSS systems to detect and distinguish objects on a highly structured sea floor, within a harbor and so forth. The reason is that SAS processing normally (but not always) enhances the along-track resolution sig-nificantly by one order of magnitude or even more compared to conventional SSS systems. Note that a SAS system has in principle a constant along-track resolution while this along-track resolution for a con-ventional SSS system decreases with across-track range. Ammunition poses a log-term hazard to the environ-ment. Many active and former military installations worldwide have ranges and training areas. Some of these include adjacent waters such as lakes, rivers or coastal water areas. Other sites for training and testing were situated on purpose at sea. In both cases military operations have led to munitions contamination. In addition, duds and lost or disposed ammunition con-taminates waters and seas during military activities.

And last but not least, a lot of ammunitions was for example dumped after WW II in the Baltic Sea and North Sea for disposal. This dumped ammunition needs to be removed e.g. before installations at sea can be set up and poses a rising significant risk to people using coastal areas for business or leisure ac-tivities due to rust. SSS systems have been used to survey areas at sea in order to image the seafloor and recover dumped WW II ammunition and to perform a risk analysis for such areas [6, 7]. Since dumped ammunition range from e.g. large ~3.5t torpedoes, over ~500kg sea mines and ~10kg shells to small bullets, the probability to detect ammunition depends to a large extent on the size of the objects looked for. For example, detecting ~10kg shells with a conventional Benthos 400kHz SIS 1625 SSS can be a difficult task due to the small size of the shells in conjunction with the limited resolution of the sonar (see Fig. 1). Only the existence of comet marks [8] created by tidal currents in the given example permits the conclusion that small objects of the size of a 10.5 cm shells are present.

Figure 1 SSS image of small objects including cur-rent induced comet marks (presumably caused by dumped ~ 10kg shells) of the coast of Helgoland [6] gathered with a Benthos SIS 1625 SSS. The given examples indicate, that objects of various sizes can be detected and located using SSS systems. Therefore, SSS imagery is often used to detect debris items and other obstructions on the seafloor that may be hazardous to shipping or to seafloor installations. In addition, the status of underwater installations like e.g. pipelines and cables on the seafloor can be inves-tigated for maintenance. A inspection survey may in-clude [9]:

- detection of burial for exposed installation or detection of exposure for buried installation,

- detection of free span of a pipeline or cable and estimation of span,

- detection of damages, - detection of buckling, - detection of debris next to the installation.

In Fig. 2 an example of an inspection task taken from [9] is given. In this example a part of approx. 2000m of a pipeline is shown from one side. This pipeline part is covered by gravel in several locations. Last but not least another application of SSS imagery shall be mentioned, which is of special importance for

Session A.1Future Security 2011 Berlin, September 5-7, 20112

areas with a long lasting history. In territories like the Mediterranean Sea maritime archaeology has derived a lot of benefit from conducting SSS surveys [10, 11].

Figure 2 SSS image of ca. 200m part of a pipeline shown from one side. The range is 0m (bottom) to 180m (top). (Image: FFI with the HISAS / HUGIN AUV [9]). 2 Synthetic Aperture Sonar

A typical SSS image shows three different image re-gions high light regions caused by an object, shadow regions behind an object and background regions caused by the sediment of the seafloor. In Fig. 3 a submarine section is shown, which illustrates this.

Figure 3 SSS image of a submarine section insoni-fied from the left side showing an object, an object shadow and the surrounding seafloor. In order to use a SSS system for detection and classi-fication or inspection, three processing tasks or deci-sions have to be taken [12]:

- Whether a pixel belongs to an object, shadow, or background region is the first task or re-quired decision for each pixel in a noisy SSS image.

- Whether an image object is of interest or a false target caused by speckle noise, stone, bot-tom structures, is the second required deci-sion for each observed image object.

- Finally, whether an object of interest is e.g. a cylinder, a truncated cone, is the final re-quired decision, which is normally regarded as classification.

The performance for each processing task depends on the physical resolution of the SSS system. One simple predictive model which describes approximately the performance of a human observer or an automatic tar-

get recognition (ATR) software analyzing SSS im-agery are Johnsons Criteria [13, 14]. According to these criteria the minimum required resolution (in the direction of the shortest object dimension and for a 50% probability to discriminate an object to the speci-fied task) is 1.5 - 3 pixels for detection, 8 10 pixels for recognition or 13 16 pixels for classification. This illustrates that the performance of the different mentioned discrimination tasks get worse if the object dimensions get small. Therefore, the physical resolu-tion needs to be as high as possible in order to avoid loses. One way to achieve an increase in resolution is to apply synthetic aperture techniques. SAS is a revolutionary development in the field of acoustic imaging of the seabed. By coherent addition of data from a series of consecutive pings a signifi-cantly longer antenna the synthetic aperture - is syn-thesized in the direction of travel. Hence, the lateral resolution is often improved by an order of magnitude or even more compared to conventional SSS systems. This results especially in a benefit for object classifi-cation. Required SAS processing steps are:

- estimation of the motion of the short physical antenna from ping to ping using data driven motion compensation (e.g. the Displaced Phased Center Array DPCA algorithm),

- estimation of the sensor path based on DPCA values without or with considering navigation information from the inertial navigation system (INS),

- synthetic aperture beam forming including fo-cusing,

- re-focusing by auto-focusing algorithms. Critical to the quality of a SAS image are position er-rors for the physical antenna along the synthetic aper-ture. Deviations from the often assumed perfectly straight path can be measured from ping to ping by data driven micro-navigation methods like DPCA and have to be corrected for. The estimate of the trajectory results by integration. One challenge is that the uncer-tainty of the trajectory increases with the number of pings. Hence, a combination of DPCA results and INS readings are normally used for correcting the position of the antenna in order to achieve the required posi-tion uncertainty of less than /16 along the synthetic aperture. This is within the (sub) millimeter range for several 10-meters of track line. In Fig. 4 the estimated lateral sway or transverse dis-placement component of the trajectory is shown for a data set gathered with an experimental SAS system attached to the SeaOtter MK II AUV in the Baltic Sea [16]. The different curves base on readings of the INS (Green) and results of the DPCA algorithm (Red). For the third curve (Blue) DPCA results for the yaw angle (small graph) were complemented by measured head-ing values from the INS. Fig. 4 illustrates, that the lat-

pipeline gravel

Session A.1Future Security 2011 Berlin, September 5-7, 20113

eral component of the sway estimate varies quite a bit depending on the chosen method to get estimates for the required navigation data. Note that the shown sway movement is only the lateral component of the trajectory. For the full trajectory of the SAS antenna the vertical sway component as well as yaw estimates and the surge estimates are also required. The question which micro-navigation method pro-vides the best results for SAS imagery can be deter-mined by using a resolution test target. It consist of a cross with five spheres of 5cm in diameter and dis-tance on each leg [16]. In Fig. 5 generated SAS im-ages for this test target and the different trajectory es-timates from Fig. 4 are shown. The best image (right) bases on a track estimate derived from the combina-tion of INS data and DPCA results. Although, some weak artifacts indicating remaining phase errors can be still noticed. Track estimates using INS data (left) or DPCA results (middle) provide much worse image results.

Figure 4 Estimated lateral sway of the SAS antenna over 80m of track line: Green based on readings of the INS, Red based on DPCA results, Blue data-driven DPCA estimates of the yaw angle

(small graph) was complemented by measured heading values from the INS.

Figure 5 SAS images of a resolution test target using track estimates derived from INS data (left), DPCA results (middle) and a combination of INS and DPCA data (right). 2.1 SAS - SAR Comparision

The idea of SAS and Synthetic Aperture Radar (SAR) is very similar. In both cases the scene echoes for a number of pings are stored along a portion of the sen-sor path. These echoes are afterwards combined co-herently by an appropriate algorithms to generate high resolution images.

In Tab. I typical system parameters (wavelength , SAR/SAS resolution D, synthetic aperture length LS, vehicle speed V) for a satellite SAR and an aircraft SAR system taken from [17] are given. These parame-ters are compared to a 200kHz SAS system with a 30 - 150m swath. Tab. I indicates that the tolerable posi-tion error r of /16 in order to avoid increased side lobe levels is system dependent. It decreases from 3.5mm for the considered satellite based SAR to 0.05 mm for the assumed 200kHz UUV-SAS system. At the same time the required position accuracy has to be ful-filled over a longer period of time t for the SAS sys-tem compared to the SAR systems, since under water it takes much longer to travel the distance of one syn-thetic aperture. The satellite needs about 0.7s to pass through one synthetic aperture of 4850m while the UUV needs about a factor of 20 more time (14s) to travel 27m. Hence, motion compensation is normally much more serious for SAS systems compared to a SAR systems [17]. System / cm D / cm LS / m v / m/s r / t Satellite ERS-1 (1991 2000)

5,7 500 4850 7000 3.5mm / 0.7sec

Aircraft SAR 5.7 50 570 200 3.5mm / 2.9sec

200 kHz UUV - SAS

0.75 2.5 27 2 0.05mm / 13.5sec

Table I: Typical system parameters wavelength , SAR / SAS resolution D, synthetic aperture length LS, and vehicle speed v for a satellite SAR and an aircraft SAR system [17] and a 200kHz SAS system. The last column gives the tolerable position error r of /16 and the time t for which this is necessary. Motion compensation is not the only difficult issue for a SAS system. Range ambiguity is also a more severe problem for sonar than radar [17]. A single or low number of element SAS avoiding range ambiguity has a very low area coverage rate which is not tolerable. Therefore, practical SAS antennas have always a multi-element receiver array with N elements and total length d (Typical N = 96 192, d = 1m - 2m). Further differences [17] are that SAR systems suffer mainly from thermal and electronic noise while SAS systems operate in addition in the noisy sea. Also sound propagation at sea is much more influenced by the environment compared to wave propagation through air or space resulting in a more unstable situa-tion for sonar. Some factors which have to be consid-ered for sonar are multipath propagation, a variable refractive index including refraction effects, temporal instabilities and sound attenuation. E.g., the SAS im-age quality may break down to less than 30% and re-cover to the normal 100% sonar range within a few ten meters of travel due to the influence of multipath propagation. This applies particularly for shallow coastal waters. Also operational conditions causing

Session A.1Future Security 2011 Berlin, September 5-7, 20114

unexpected UUV movement may significantly de-grade a SAS image. E.g. cross-currents may lead to bearing angles or surface waves may result in rapid changes in the altitude of the UUV. Both causes blurred SAS images. Therefore, it is essential to un-derstand all limiting factors in detail in order to use the SAS technology successfully worldwide under dif-ferent environmental conditions. 3 Image Processing

After SAS processing the next challenge is the analy-sis of the obtained images. A SAS system typically produces about 500 000 pixels per second. Due to this amount of visual information, an operator needs sup-port on the job. Object detection can be done, for ex-ample, by computer-aided detection and classification algorithms, which highlight and enlarge image regions with suspicious objects. A second approach, which is especially of interest for military applications is Au-tomatic Target Recognition (ATR). An ATR system analyses SAS or SSS imagery completely autonomous in order to detect and classify objects on the fly during the mission. The main military benefit is mission time which is reduced to a half or more by ATR. Since an operator is absent in the decision chain on board of an UUV, requirements for ATR systems are much higher than for operator assistance. The image processing system under development at WTD 71 consists of several processing steps includ-ing pre-processing, screening (detection), reduction of false positives and classification [19, 20]. We have recently complemented our activities by addressing special image processing issues such as the influence of image resolution [13], edge preserving filtering [21] or fusion [22, 23, 24] on the outcome of the en-tire processing chain. We use in parallel different al-gorithms for each processing step, to obtain a robust system. For physical reasons, side-scan sonar images are very noisy. Hence goal of the required noise filtering is to improve the image quality as much as possible by noise reduction and at the same time preserve the structural features and contours (e.g., object edges) and textural information of the image. Several filters can be used for this task and have been investigated like Median -, Kuwahara -, Bilinear -, Curvelet - or UINTA - Filter. In Fig. 6 an example is given. The im-portance of pre-processing in total, which affects all subsequent processing steps significantly, is often un-derestimated. Screening is the second step in our ATR processing chain. It allows to detect image regions, that contain a potential object. Since this processing step must be fast and robust, typically a high number of image re-gions are also highlighted which contain false targets. Screening algorithms may base on statistical features,

contour features, correlation features or other proper-ties of the image. The algorithms are typically opti-mized for SSS or SAS data. As an example, a screen-ing result for SAS data is shown in Fig. 7. False targets can be caused during screening e.g. by seabed structures or speckle noise. Since screening reduces the volume of the remaining image data sig-nificantly, computationally intensive algorithms, such as Active Contour approaches (Snakes), can be used for the reduction of false targets. Snakes minimize an energy function associated to the contour of an object and / or object shadow. How well a snake matches a conture, a snake sepertes image regions with different statistical properties or a snake represents an assumed model are some factors, which can determine the en-ergy function. As an example, image results of an Gradient Vector Flow Snake taken from [20] are shown in Fig. 8.

Figure 6: Illustration of edge-preserving behavior of a Fast Bilateral Filter realized by Fraunhofer IOSB. The original SAS object image, the filter result, and the difference of both as an indicator of the loss of con-tour information are shown.

Figure 7: SAS image of the MUSCLE system (NURC) with five known objects, highlighted by computer-aided detection. Each detected image region with a potential object is overlaid in a filtered and close-up version. Afterwards commonplace objects such as stones are separated from interesting objects such as sea mines during classification. This is the most challenging processing step since it requires high physical resolu-tion [13] and the response of an object located on the seafloor may vary significantly depending on e.g. ob-ject position, sonar design / sonar parameters and en-vironmental factors (e.g. ripple, vegetation).

Session A.1Future Security 2011 Berlin, September 5-7, 20115

Figure 8: Result of Gradient Vector Flow (GVF) Snake algorithm for SAS data of a truncated cone and a cylinder gathered by the MUSCLE system (NURC). The results of the edge-preserving UINTA filter as precursor and the Snake in the original images are shown on the left and right, respectively. The classification process is divided into two main steps, feature extraction and association of these fea-tures to a certain class of objects. Features represent condensed information of an object image and can range from simple structures such as points, corners or edges to more complex structures such as shapes or ridges. Mathematical or statistical values may also be used as features. Feature extraction from an ROI results normally in a n-dimensional vector of numerical values that repre-sent an object. This vector may be reduced in dimen-sion and is then assigned to a certain object group, which represents objects with similar attributes (e.g. cylindrical objects). This is the actual classification step. We use a Probabilistic Neural Network (PPN), K-Nearest neighbor (KNN) and two forms of Support Vector Machine (SVM) algorithms for this step.

Figure 9: Multi aspect fusion results for different combinations of aspect angles Finally, studies [22, 23, 24] indicate that fusion is a promising strategy to further increase the probability of detection and decrease the number of false posi-tives. In Fig. 12 the fusion gain in a multi-view framework, where a set of acoustic images for differ-ent aspect angles are processed, is shown. Object im-ages are processed individually down to classification. Afterwards the results for three different aspect angles

are fused using a voting strategy and compared to the results for a single view. This results in a significant improvement for correct classification especially in the region of low false alarms. 4 Conclusion

In the first chapter different application were de-scribed. Main civil applications for SSS imagery in-clude search and rescue missions ranging from crashed airplanes to dumped ammunition and inspec-tion missions. The capability to detect, recognize and classify some-thing in a SSS image depends on the resolution of the measuring SSS system. This explains the importance of SAS processing. The performance of high resolu-tion SAS processing is very sensitive to the environ-mental and operational conditions. This motivates cur-rent research activities related to SAS. Due to the high amount of visual information an op-erator needs support analyzing SSS / SAS imagery. First versions of computer aided detection and classi-fication software are available mainly for military purposes. Currently Automatic Target Recognition software, which can analyze SSS / SAS imagery fully autonomous, is in the focus of scientific research.

Acknowledgement

In recent years the Research Department of WTD 71 has investigated image processing methods for object detection and classification in conventional side scan sonar and synthetic aperture sonar images in coopera-tion with the FU-Berlin and Fraunhofer IOSB (former FGAN-FOM). The purpose of these activities was to implement computer aided detection and classification methods in order to investigate and improve these al-gorithms and to test all implemented algorithms based on measured image data. Recently these activities have been complemented by investigating SAS proc-essing in cooperation with NURC and Atlas Elek-tronik. The authors would like to thank NURC for supporting the SAS activities at WTD 71 FWG and for providing SAS data from the COLOSSUS 2 sea trial in the Baltic Sea with the MUSCLE system.

References

[1] http://en.wikipedia.org/wiki/Side_scan_sonar [2] http://www.ifm-geomar.de/ Wracksuche AF447 -

MV Alucia [3] History and Honor news article, 7 Sep. 2010

http://www.mod.uk/DefenceInternet/Defence News/HistoryAndHonour/

[4] http://www.edgetech.com/edgetech/gallery/ item/4125-p-side-scan-sonar-system

Session A.1Future Security 2011 Berlin, September 5-7, 20116

[5] http://www.l-3klein.com/?page_id=17 [6] H. Fiedler und S. Behringer, Sonaruntersuchun-

gen in der Helgolnder Tiefen Rinne, Wehrtech-nischer Bericht WTD 71 - 0027/2010 WB, Eck-ernfrde 2010

[7] http://www.schleswig-holstein.de/AFK/DE/ Das Amt fr Katastrophenschutz, Munitionsalt-

lasten im Meer, Regionale Informationen, Helgo-land

[8] F. Werner, G. Unsld, B. Koopmann, A. Stefa-non, Field Observation and Flume Experiments on the nature of Comet Marks, Sedimentary Ge-ology 26, 1980, p. 233-262

[9] T.O. Sb, Seafloor Depth Estimation by means of Interferometric Synthetic Aperture Sonar, PhD Thesis Sep. 2010, University of Troms, Norway and T.O. Sb, H.J. Callow and P.E. Hagen, Pipeline inspection with synthetic aper-ture sonar, Proc. 33th Scandibavian Symposium on Physical Acoustics, 07 10 Feb. 2010

[10] R. Quinn, M. Dean, M. Lawrence, S. Liscoe, D. Boland, Backscatter responses and resolution considerations in archaeological side-scan sonar surveys: a control experiment, J. Archaeological Science 32, 2005, p. 1252 1264

[11] European Commission - ITC Research in FP 7: VENUS - Virtual ExploratioN of Underwater Sites, Final Report, 2009, http://www.venus-project.eu/

[12] Advances in Sonar Technology, Edited by S.R. Silva, 2009 In-teh, ISBN 978-3-902613-48-6

[13] F. Langner, C. Knauer, W. Jans and A. Ebert, Side Scan Sonar Image Resolution and Auto-matic Object Detection, Classification and Iden-tification, Proc. IEEE Oceans09, Bremen, Ger-many, 11 14 May 2009.

[14] J. Johnson, Analysis of Imaging Forming Sys-tems, Proc. Image Intensifier Symposium, AD 220|60, p. 244 - 273, Warfare Electrical Engi-neering Dept., US Army Engineering Research and Development Laboratories, Ft. Belvoir, VA, 1958. Reprint in: R.B. Johnson, R. B. and W.L. Wolf (eds.), Selected Papers on Infrared Design, SPIE Proceedings vol. 513, pp. 761 - 781, 1985.

[15] R. Heremans, A. Bellettini, M. Pinto, Milestone: Displaced Phase Enter Array, Sep. 2006, http://www.sic.rma.ac.be/~rhereman/milestones/dpca.pdf

[16] J. Rademacher, Interferometry performance - shallow water experiments in the Baltic Sea, UAM 2011, Kos, Greece, 20 24 June 2011.

[17] S. Holm, Synthetic Aperture Radar and Sonar SAR and SAS, Department of Informatics, Uni-versity of Oslo, 2010, http://www.uio.no/

[18] P.T. Gough and M.P. Hayes, Ten key papers in synthetic aperture sonar, Proc. Acoustics08, Pa-ris, France, 29 June 4 July 2008 and M.P. Hayes and P.T. Gough, SYNTHETIC APER-TURE SONAR: A MATURING DISCIPLINE,

Proc. ECUA2004, Delft, The Netherlands, 5 8 July 2004

[19] F. Langner, C. Knauer, W. Jans and W. Middel-mann, Image processing in Side Scan Sonar Im-ages for Object Detection and Classification, Proc. UAM2009, Nafplion, Greece, 21 26 June

[20] F. Langner, W. Jans, C. Knauer and W. Middel-mann, Computer Aided Detection of MLOs in Side Scan Sonar Images, Proc. UDT Eu-rope2010, Hamburg, Germany, 8 10 June 2010

[21] W. Jans, F. Langner, C. Knauer and W. Middel-mann, The effect of pre-processing on the out-come of an CAD/CAC system for underwater objects in SAS and conventional side scan im-ages, Proc. ECUA2010, Istanbul, Turkey, 5 9 July 2010.

[22] F. Langner, C. Knauer, W. Jans and A. Ebert, Performance gain by fusing classification results for different aspect angles in SAS side scan im-ages, Proc. ECUA 2010, Istanbul, Turkey, 5 9 July 2010.

[23] F. Langner, W. Jans, C. Knauer, and W. Middel-mann, Benefit for screening by automated acous-tic data fusion, Proc. UAM 2011, Kos, Greece, 20 24 June 2011.

[24] F. Langner, W. Jans, C. Knauer, A. Ebert, Benefit for classification by automated acoustic data fu-sion, UAM 2011, Kos, Greece, 20 24 June 2011.

Session A.1Future Security 2011 Berlin, September 5-7, 20117

Detection of High Power Microwaves Christian Adami, Christian Braun, Peter Clemens, Hans-Ulrich Schmidt, Michael Suhrke, Hans-Joachim Taen-zer, Fraunhofer-Institut fr Naturwissenschaftlich-Technische Trendanalysen INT, Germany Yolanda Rieter-Barrell, TNO, The Netherlands

Abstract

The growing threat to critical infrastructure by high power microwaves (HPM) also increases the importance of detection facilities for electromagnetic fields with high field strength. We discuss HPM detection principles as well as capabilities and limitations of existing HPM detectors. Then we describe the basic requirements for a system for the detection and identification of HPM threat signals and a demonstrator of a single-channel HPM detection system for mobile and stationary use. The system allows the measurement of amplitudes within a very high dynamic range, the pulse width, pulse repetition frequency and the number of pulses.

1 Introduction

The availability of components to build low-tech high power microwave (HPM) sources together with the increasing dependence on electronic devices and sys-tems has lead to a situation where all microprocessor controlled electronics can be disabled with HPM at-tacks at least temporarily with medium sized device within distances from several 10 m to a few hundred meters. This is crucial all the more due to the reliance of critical infrastructures on electronics. The exam-ples of vulnerable systems range from commercial IT electronics and network equipment used also in the military area and in civilian security applications as commercial off the shelf (COTS) electronics to elec-tronic systems in vehicles, surveillance equipment and logistics. Because of their easy availability, it is very likely that also persons or groups with criminal or terrorist inten-tions can acquire such HPM systems. These then could be used for burglaries, raids, blackmails and attacks in cases where electronics is responsible for the safety of persons and property. Without detection and alarm systems it is easy for attackers to test their HPM devices without being discovered. For this rea-son failures and malfunctions of own electronic sys-tems cannot be traced back to an electromagnetic at-tack. This is the case the more so because of the gen-eral lack of awareness of the electromagnetic threat. Therefore, it becomes increasingly important to de-velop and investigate detection techniques for this threat. For intentional electromagnetic interference (IEMI) mainly the following procedures come into question: Pulsed radio frequency (RF) emissions (narrow-

band sources), most conveniently at frequencies be-

tween 30 and 3000 MHz, with pulse widths of about 0.1 to 10 s.

Single or repetitive ultra-wideband pulses (UWB) with rise times and pulse widths in the range 10 ps to 1 ns.

Single and repetitive broadband pulses, maybe also damped sinusoidal (DS) signals, etc.

Continuous wave (CW) RF emissions in the lower GHz range.

At the target device electromagnetic field strengths must be generated which are sufficiently far above the immunity of the unit. For CW signals and digital elec-tronic devices the necessary field strength is roughly in the range above 100 V/m, for pulsed RF or micro-wave signals and damped sinusoidal oscillations ap-proximately 1000 V/m and for very narrow, steep-edged pulses a few kV/m. The paper gives a short overview of detection princi-ples and previous detector development in the second section. Section 3 describes development at Fraun-hofer INT of a demonstrator of a detection system with high amplitude dynamics based on logarithmical amplifier/detector ICs which is able to cover a large frequency range with broad-band antennas and can be deployed stationary or vehicle mounted. The paper concludes with summary and outlook.

2 HPM Detection: Principles and Previous Developments

2.1 Detection Principles

In the past, basically only low-impedance broad-band diode detectors have been available as actual detec-tion elements as both high-impedance diode detectors and thermal power meters have much too large re-

Session A.1Future Security 2011 Berlin, September 5-7, 20118

sponse times for short pulses and are useful only for the recording of time averaged signals. The disadvan-tage of these detectors is the limited amplitude dy-namics for pulse measurements which strongly limits the detectable range of amplitudes and consequently the possible detection range [1,2]. The voltage vs. power characteristics of a typical Schottky diode de-tector head [3] gives an achievable dynamic range of about 20 to 25 dB under the realistic assumption that pulsed voltages lower than 1 mV can not be identified with a digital oscilloscope. This can be hardly in-creased even with low-noise signal preamplifiers. The measurement of the transient response of such a Schottky diode detector shows that it is possible to resolve rise and fall times of some nanoseconds [3]. Special detectors as resistive sensors based on the electron heating effect in semiconductors in a strong electric field or lithium niobate crystals utilising the electro-optical effect are mainly useful for high field strengths [1,4,5]. Recently, a number of highly broad band and at the same time in part relatively inexpensive logarithmical amplifier/detector ICs have entered the market which allow to avoid the disadvantages of diode detectors. In principle, these devices consist of a large number of linear broadband amplifiers with defined gain, which are connected in series. At the output of each amplifier is a linear diode detector. The output signals of all detectors are added via an analogue summing circuit, so that a quasi-logarithmic detector character-istic is achieved. The measurement of the detector characteristic of such a logarithmic amplifier/detector module shows a dynamic range above 60 dB [3], which represents a significant improvement over the previously used diode detectors. The frequency spans from 1 MHz to 8 GHz with an amplitude correction of a few dB above 5 GHz. Also the rise and fall times of the detected signals meet the requirements for HPM detection completely [3]. With such characteris-tics the properties of these ICs considerably surpass some of the conventional logarithmic ampli-fier/detector units as employed e. g. in radar warning systems.

2.2 Previous HPM Detector Develop-ments

For surveillance of the surroundings of electronic fa-cilities against electromagnetic attacks detection sys-tems are needed, which register at least the occur-rence of HPM signals as such. Those devices, option-ally enhanced by a coarse display of amplitude levels and number of threat pulses, are in many cases suffi-cient as pure alarm units. Such small-sized low-cost systems already have been realised in different im-plementations in form of battery-powered pocket or hand-held units with integrated omni-directional broad band antennas [1].

As a first example Canary is a prototype sensor de-signed and developed by Qinetiq, United Kingdom. According to the Canary datasheet [6] the specifica-tions of the detector are as follows: Signal types: HPM, High Altitude Electromagnetic

Pulse (HEMP), Non-Nuclear Electromagnetic Pulse (NNEMP), DS (for repetition rated waveforms), UWB (for repetition rated waveforms, minimum pulse width detected: ~ 300 ps).

Frequency range: 10 MHz - 8 GHz (calibrated), up to 40 GHz has been detected.

Sensitivity threshold: Low level: 1 mW/m2 (Eeff ~ 1 V/m). High level: 1 W/m2 (Eeff ~ 20 V/m). Electromagnetic Pulse (EMP): 1 kW/m2, single

pulse (Eeff ~ 615 V/m). Maximum input level: not known. The sensitivity levels can be tailored to meet require-ments for specific applications. The LO, HI and EMP detection levels of 1

mW/m2, 1 W/m2 and 1kW/2 correspond to an effec-tive field strength (Eeff) of 0.6 V/m, 19.4 V/m and 614 V/m, respectively. According to datasheet [6] these levels are based on the following: LO warning; Low alarm threshold indicating that

an EM event has been detected of sufficient mag-nitude to cause IT upset or degradation (indication visible and audible).

HI warning; High alarm threshold indicating that an EM event has been detected of sufficient mag-nitude to cause IT prolonged disruption or damage (indication visible and audible).

EMP warning: Indication that a single EMP event has occurred. Description of the physical operation (indication visible and audible).

The second example is the microwave microphone, a first generation high power microwave detector de-signed and developed by Market Central, USA. Ac-cording to the product sheet the specifications of the detector is as follows [7]: Signal types: Transient Electromagnetic Device

(TED), UWB and CW. Frequency range: 900 MHz - 2.9 GHz (flat re-

sponse), 400 MHz - 3 GHz (-10 dB). Sensitivity threshold: ~ 100 V/m. Visual indication via a 10-segment LED covering a

30 dB range (up to ~ 3 kV/m). Maximum input level: not known. The detector can sense signals in all three polarisation axes and has the capability to indicate detection real-time or in a peak hold mode. It has an audible alarm with a false alarm indication. The internal antenna can be replaced by external antennae optionally [7]. The detector contains a rechargeable battery and USB connectivity. USB connectivity is presently used to recharge the battery, but will be used for networking and remote reporting at a later stage. The detector is relatively light in weight and consumes very little power [7].

Session A.1Future Security 2011 Berlin, September 5-7, 20119

The third example is a high power microwave detec-tor prototype designed and developed by TNO, The Netherlands [8]. The detector has been designed to meet specifications on signal types, flat frequency re-sponse, omni-directionality and response time. More-over, the detector was designed to be low-cost. Less attention was paid to the maximum input level the de-tector is resilient too. The characteristics of the detec-tor are as follows: Signal types: Carrier-based pulses, UWB and CW. Frequency range: 100 MHz - 8 GHz (flat response). Sensitivity threshold:

Low level: 3 V/m. Medium level: 10 V/m. High level: 40 V/m

Visual indication via 3 LEDs, separate indication CW and carrier-based pulses.

Maximum input level: > 1000 V/m (carrier based), > 3 kV/m (UWB).

The detector has shown to have a flat frequency re-sponse in the frequency range tested. The detector is fast enough and sensitive enough to detect CW, car-rier-based and UWB pulses in all directions (in cur-rent set-up only one polarisation axis). The maximum input level is not stated in the specifications but has been further researched [9]. Sensitivity and robustness tests of some of the exist-ing detector prototypes show that developers should take special care of robustness of detectors against HPM [9].

3 HPM Detector for Mobile and Stationary Use

3.1 Detector Requirements and Devel-opment Concept

For permanent surveillance of high-value or mission-critical stationary facilities and especially for search and identification of HPM sources such devices should, however, feature an extended scope of per-formance. These additional features include the dis-play of field strength of the threat signal, an ampli-tude dynamics (i. e. detection range), which should be as large as possible, the counting of pulse number or display of pulse repetition rate, the display of pulse width, preferably a frequency independent display of field strength in a wide frequency range, and finally the directional and polarisation independence of the receiving antenna or a radiation pattern with defined wide angular range of constant gain for sector surveil-lance. The detectors should be able to detect pulsed electromagnetic fields with threat field strengths above 1 kV/m independently of frequency, be im-mune to field strengths of some 10 kV/m and be able to detect HPM sources at medium distance (i. e. field strengths down to at least 100 V/m). The detection of all signal types from CW via narrow band and

damped sine pulses to ultra wide band (UWB) pulses ideally requires response times in the upper picosec-ond range. The classification of detected incidents ac-cording to pulse form, i. e. amplitude, pulse width and pulse number or repetition rate can help to identify false alarms. Further development stages should realise a coarse to medium display of the direction of HPM impact for localisation of an attacker and of the HPM frequency via filter benches. The devices should be operable with batteries for a certain time in addition to a sta-tionary or on board power supply for operation from vehicles. The detection system developed at Fraunhofer INT meets the following criteria in the first phase: Notification that a pulsed electromagnetic field was

detected independent of frequency with threat level field strength (> 1 kV/m).

Damage immunity against field strengths of up to several 10 kV/m.

Frequency independent detection of HPM sources in medium distances (e.g. detection of E > 100 V/m) for warning and searching.

Measuring dynamics preferably > 60 dB. Polarization independence. Directional independence (at least in the horizontal

plane) or in a defined sector (e.g. 90 degrees). Classification of the detected events by amplitude,

pulse duration, pulse repetition frequency, form, etc.

In the first stage the system is built as a single-channel assembly with a polarization-independent broadband antenna and a logarithmic ampli-fier/detector module. To stay within the usable range of the amplifier/detector module an attenuator of 60 dB is necessary between antenna and amplifier input. A fast PIN diode limiter is connected at the entrance of the IC in order to avoid damage of the device even in the worst case. Figure 1 shows the block diagram of the system. Accordingly values for field strength can be obtained between < 100 V/m and > 10 kV/m.

DISCR . C OU NTER

LOG. -AMPLIFI ER -DETECTOR

-60dB 0...-60dBm

GAIN~ 0dB

P(an t)

DIGITAL-OSCIL LOSCOPE

ATTEN-U ATOR

SPIR AL-ANNTEN NA

DI OD E-CL IPPER

Figure 1 Block diagram of the first development stage of the HPM detection system

The signal amplitude and pulse shape are measured with a high-speed digital oscilloscope, which in turn is read out by a PC via GPIB. In addition to the

Session A.1Future Security 2011 Berlin, September 5-7, 201110

graphical representation of the envelope of the HPM pulse the evaluation software shows amplitude and field strength, pulse width and pulse repetition rate. Furthermore, the number of the threat pulses can be registered by a separate counter or by counting the trigger events of the oscilloscope.

3.2 Demonstrator of a Single Channel HPM Detection System

Figure 2 shows the layout of the demonstrator. One recognizes the spiral antenna on the left, the battery-powered multi-channel oscilloscope for signal proc-essing and the RF part in the middle and the computer with GPIB interface and the necessary analysis and display software in the right. The shielded RF part contains the logarithmic amplifier/detector module and an appropriate input circuit for self-protection.

Figure 2 Demonstrator of overall system

The spiral antenna is designed for the frequency range 0.5 - 2 GHz but can also be used between 0.5 and 10 GHz without significant gain changes and has a direc-tional pattern with a width of greater than 90 degrees (-1 dB to -2 dB). Due to the circular polarization the spiral antenna receives linear polarized signals inde-pendent of the polarization plane. The system can be operated by mains or on-board power supply and with internal batteries, respectively. For this reason the oscilloscope has a built-in battery, from which also the RF part can be supplied. The computer can also run with its internal battery. In this way the detection system can be flexibly moved around, without relying on an external power supply. The antenna and the RF part can be placed outside protected zones. At expected high field strengths the oscilloscope and the computer should be primarily used within a shield. In the search mode at low field strengths the screening of a motor vehicle should be sufficient.

3.3 First Tests of the HPM Detection System

Tests of the detector have been carried out in the TEM waveguide of Fraunhofer INT using pulsed HPM sources with frequencies between 150 MHz and 3.4 GHz and a pulse width of 1 s and low power sources up to the upper frequency 8 GHz of the TEM waveguide. The field strength has been measured and compared with the detector characteristic in the meas-urement setup as in Figure 3. The test setup with the spiral antenna in front and the shielded RF part of the detection unit behind it in the TEM waveguide is shown in Figure 4. The digital oscilloscopes and the controlling computer were positioned outside the shielded hall.

HPM detector S

EH

Pulsed source

Directional coupler

.

TEM waveguide

Digital oscilloscope Digital oscilloscope

GPIB

Controlling computer

Udet

Diode detector

Amplitude

Figure 3 Measurement setup for detector tests

Figure 4 Test setup in TEM waveguide

The detector characteristic was calculated considering the measured antenna factor of the spiral antenna and the 60 dB attenuator besides the characteristic of the logarithmic amplifier/detector module itself from a minimum frequency of 500 MHz limited downward by the antenna factor to 8 GHz as maximum fre-quency of the logarithmic amplifier/detector module. In parallel, the field strength was measured using a directional coupler and a diode detector to determine the input power into the waveguide (cf. Figure 3). The latter is related to the field strength at different measuring points inside the waveguide by calibration

Session A.1Future Security 2011 Berlin, September 5-7, 201111

measurements. As an example, Figure 5 shows the comparison between measured and calculated results for the frequency f = 1.2 GHz. Field strengths be-tween some 100 V/m and about 1.5 kV/m have been generated at the chosen measuring point inside the waveguide for this frequency. Finally, Figure 6 shows the frequency dependence of the HPM detector characteristic determined at con-stant detector voltage for frequencies between 500 MHz and 8 GHz with E0 = E(f = 1.0 GHz). The sensi-tivity of the detector, derived from the necessary field strength to obtain a certain detector voltage, decreases with increasing frequency. One reason for that is the frequency dependence of the antenna factor at con-stant gain, another reason is the frequency depend-ence of the gain itself.

1

10

100

1000

10000

0 0,1 0,2 0,3 0,4 0,5|Udet[V]|

E[V

/m]

calculatedmeasured

Figure 5 Comparison of measured and calculated results for f = 1.2 GHz

-5

0

5

10

15

20

25

30

0 1 2 3 4 5 6 7f[GHz]

E-E

0 [dB

]

8

Figure 6 Frequency dependence of detector charac-teristic

So far, tightness and robustness (electromagnetic im-munity) of antenna and RF part of the HPM detector (cf. Figure 4) have been tested against field strengths up to 1.5 kV/m for frequencies between 150 MHz and 3.4 GHz. In the future, robustness tests will be carried out also in the reverberation chamber for field strengths above 10 kV/m.

4 Summary and Outlook

For security reasons it is very important to have avail-able a detection system to protect critical equipment, systems and infrastructure against IEMI threats. For this purpose we have given an overview of HPM de-tection principles as well as capabilities and limita-tions of existing HPM detectors. Sensitivity and ro-bustness tests of some representatives show that ro-bustness of detectors against HPM is an important issue. We have described the basic requirements for a system for the detection and identification of HPM threat signals and have discussed single-channel dem-onstrator of an HPM detection system at Fraunhofer INT. In its first stage it enables the detection and iden-tification of HPM threat by measuring the field strength within a very high dynamic range, the pulse width, pulse repetition rate and the number of pulses. In future a number of multi-channel detection systems will be developed at Fraunhofer INT on the basis of the single-channel demonstrator. The first step will be a rough determination and display of the direction of the incident threat pulse. For this purpose four detec-tion channels can be used with four spiral antennas and logarithmic amplifier/detector modules. As angu-lar ranges of 90 degrees can be covered with the pre-viously described antenna types the detection in the entire 360 degree azimuth range is possible with four antennas. The direction of incidence can be roughly evaluated and displayed by analyzing the amplitudes in the four channels. The azimuthal information can be determined in more detail through the use of an-tennas with narrower antenna diagrams and the in-crease of the number of channels. To obtain more precise information about the carrier frequency of the threat pulse, the signal of an antenna can be split to a larger number of channels via a coax-ial power divider, whose outputs are each equipped with a band pass filter. Depending on the bandwidth of the filter and number of frequency channels the carrier frequency can then be determined from the amplitude ratios of the individual channels. This method was used in principle already in [2] for a rough determination of the unknown frequencies of pulsed microwave signals, but at that time with a much smaller dynamic range compared with modern logarithmic amplifier/detector modules. In the demonstrator systems discussed up to now only the antenna and the RF part can be exposed to the full threat field strength, whereas the signal acquisition and processing with a digital oscilloscope and a port-able PC have to be operated in a shielded environ-ment in most cases. In a further expansion stage the analogue-digital conversion and analysis of the pulses could also be conducted in the highly shielded RF module instead by an external oscilloscope. The read-out of the computer implemented in the RF module would then be carried out over fibre optic cables to an also highly shielded control unit. In this way the over-

Session A.1Future Security 2011 Berlin, September 5-7, 201112

all system will be usable in any environment without restrictions. Such an HPM detection system is flexible enough to be used vehicle mounted but also stationary for pro-tection of fixed critical infrastructure installations. It complements the necessary shielding and protection measures and alternatively the hardening of critical infrastructure components against high power micro-waves.

References

[1] Rieter-Barrell, Y.: Review on HPEM Detection techniques, Final Report of NATO RTO SCI-198 Task Group on Protection of Military Networks against High Power Microwave At-tacks, to be published

[2] Braun, Ch.; Clemens, P.; Schmidt, H.-U.; Taen-zer, H.-J.: Ein Mehrkanal-Mikrowellen-Spektro-meter zur Messung einmaliger Mikrowellenfel-der im Frequenzbereich 1 26 GHz. Fraunhofer INT Arbeitsbericht IB-7/00, 2000

[3] Adami, Ch.; Braun, Ch.; Clemens, P.; Schmidt, H.-U.; Suhrke, M.; Taenzer, H.-J.: HPM Detec-tion System for Mobile and Stationary Use. Demonstrator of a Single-Channel Device, Fraunhofer INT Arbeitsbericht IB-08/11, 2011

[4] Dagys, M.; Kancleris, .; Simnikis, R.; Schami-loglu, E.; Agee, F. J.: Resistive Sensor: Device for High-Power Microwave Pulse Measure-ment, IEEE Antennas & Propagation Magazine, Vol.43, No 5, 64-79, 2001

[5] Braun, Ch.; Suhrke, M.; Taenzer, H.-J.: Testing of Resistive Sensors at INT EME Facility, Fraunhofer INT Arbeitsbericht IB-18/10, 2010

[6] Qinetiq. Datasheet Canary EM detector; issue 3 [7] Market central. Product sheet Microwave mi-

crophone [8] TNO: User manual HPM detector, 2009 [9] Adami, Ch.; Braun, Ch.; Clemens, P.; Rieter-

Barrell, Y.; Schmidt, H.-U.; Suhrke, M.; Taenzer, H.-J.: Detection of High Power Microwaves, NATO RTO SCI-232 Symposium on High Power Microwaves and Directed Energy Weap-ons, Virginia Beach, USA, 2011

Session A.1Future Security 2011 Berlin, September 5-7, 201113

Clinotrons High Power Sources for Terahertz Sensors D. M. Vavriv1, K. Schuenemann2, V. A. Volkov1, and A. V. Somov1

1 Institute of Radio Astronomy of the National Academy of Sciences of Ukraine, 61002 Kharkov, Ukraine

2 Technical University Hamburg-Harburg, Hochfrequenztechnik, D 21071 Hamburg, Germany

Abstract

A series of clinotron tubes and clinotron based oscillators, which effectively operate throughout the millimeter and submillimeter wavelength bands, has been developed at the Institute of Radio Astronomy. In this paper, the design and characteristics of these devices as well as prospects of clinotron oscillators for improving their char-acteristics and extending the operation frequency range are considered. An example of using a 300 GHz clino-tron oscillator in a beam steering imaging system is presented.

1 Introduction The THz frequency range is becoming increasingly attractive for researchers due to a number of current and potential scientific and practical applications of such frequencies. It is clear now that the development of this frequency range will have a dramatic impact on remote sensing, telecommunication, radio astron-omy, plasma diagnostics, medical imaging, security screening, industrial-process monitoring, monitoring of atmospheric pollutions, spectroscopy and many other areas [1-5]. The main factor that still essentially prevents the application of THz-frequencies is related to a lack of suitable sources operating within this range. Most of the applications call for compact, high-power, tunable, room-temperature oscillators. At pre-sent time, sources for terahertz radiation are direct multiplier-based sources, frequency mixers, the gyro-tron, the backward wave oscillator (BWO), the far in-frared laser, optically pumped lasers, free electron la-sers, synchrotron type sources, single-cycle sources, and some others. Each of these sources has its own advantages and disadvantages, but none of them meets completely the formulated requirements. As for the low-frequency part of the THz-region - up to about 3 THz, the BWO [6, 7] can be considered as the most suitable candidate for wide practical applica-tions. The only disadvantage of this type of oscillators is related with its low power. In this presentation, we describe a THz-tube, called the clinotron, which can eliminate the indicated shortcoming of the BWO. The clinotron was invented by Ukrainian scientists [810], and it has already proved its efficiency for frequencies up to 500 GHz. In the next section, design features of the clinotron and its main characteristics are de-scribed. A high level of the output power and a wide tuning frequency range of the clinotron tubes make

them attractive for the development of various imag-ing systems for security applications. It has been al-ready demonstrated that imaging at submillimeter wavelengths offers the definite advantage of being able to scan objects through various different mate-rials [11]. For instance, in security implementations, dangerous objects can be identified through clothing, cardboard packaging, and other concealing materials. In aviation and flight safety, submm-wave imaging can be used to provide vision through fog, haze, dust storms, and other harsh weather conditions during air-craft landing approach or contour chasing [12]. Vari-ous approaches have been already proposed for the development of such imaging systems [13, 14]. For example, a beam steering technique is typically used to focus scanned area segments onto a single of mul-tiple receivers [13]. In Section 3, we describe a re-cently developed beam steering imaging system that is based on a 300 GHz clinotron oscillator with an output power level of about 200 mW. Such system offers advantages related with a long operation range.

2 Clinotron oscillators The clinotron is similar to the BWO in the sense that it utilizes the interaction of an electron beam with spa-tial harmonic components of the electromagnetic field of a slow-wave structure. However, some essential modifications in the tube design were introduced. The distinguishing features of the clinotron are the follow-ing: (i) The electron beam is inclined to the surface of a grating as illustrated in Fig. 1, where the schematic of the tube is given. By varying the tilt angle , it is easy to optimize the length of the effective interac-tion space without changing the geometry of the tube.

Session A.1Future Security 2011 Berlin, September 5-7, 201114

(ii) The beam thickness is large compared to that in the conventional BWO. (iii) The electrons are bunched in an exponentially growing field. (iv) A wide electron beam is used, which enables to increase the beam current and the output power simul-taneously. (v) The clinotron is usually built as a resonance de-vice. Nevertheless, the electronic tuning range of the operating frequency is large what is achieved by suc-cessively exciting the resonator modes with beam voltage variation. The above approaches have resulted in the develop-ment of clinotron tubes for the frequency range from 30 GHz to 500 GHz [8, 9]. The output power level of these tubes is at least an order of magnitude larger than for conventional BWOs [6, 7]. For example, 300 GHz and 500 GHz tubes provide a power level of about 500 mW and 100 mW, respectively. Physical dimensions, weight, and the operating voltage of cli-notrons are comparable with those of BWOs. The en-ergy output in the clinotron is arranged by means of a waveguide, as is schematically shown in Figure 1.

Figure 1 Configuration of the beam, the grating, and the cavity in the clinotron

However, tubes with a distributed energy output, where the energy is directly radiated via a transparent window, as shown in Figure 2, have been developed and produced as well [8-10]. The output energy is coupled out here by means of the Smith-Purcell radiation. Thus such clinotron tubes ac-quire the circuit solution used in orotron oscilla-tors [15]. In the clinotron, this radiation is organized by using a two-periodic grating. Such clinotron tubes are easily matched with quasioptical transition lines and, therefore, this modification is especially interest-ing for applications in THz-systems. Due to their unique characteristics, the clinotron tubes have already been used in various electronic systems like plasma diagnostic instruments, short-range ra-dars, local oscillators, etc. To meet the above applica-tions, compact, solid-state, high-voltage power sup-plies have been developed for clinotron-based oscilla-tors [16]. These supplies feature both high efficiency and reliability. The possibility of the development of

clinotron-based synthesized oscillators with relative frequency stability of about 10-7 has already been demonstrated, too.

Figure 2 Schematic view of the clinotron with dis-tributed energy output

The theoretical studies [17-19] indicate that the clino-tron has a large potential for further increasing both the operating frequency and the output power. In or-der to implement this potential, clinotron tubes with a denser and more intensive electron beam should be realized. It is important that the space charge and temperature effects do not impose serious limitations on an essential current density increase as compared with those values used in the present clinotron design. According to our recent simulation results, output power levels of about 2 W and 70 W can be achieved at frequencies around 1 THz in CW and pulsed oper-ating modes, respectively. To embody such tubes, the beam current density should be increased to about 100 A/cm2 for the CW mode and to about 1000 A/cm2 for the pulsed mode. The beam cross section can be 0.05 mm x 2.5 mm, which is the same as in the already produced tubes.

3 Clinotron-based imaging system

A block diagram of the clinotron-based imaging system that has recently been developed and tested is shown in Figure 3. We developed an imager design that produces a raster of the scene with a minimal loss of the radiated energy in order to utilize the high out-put power capability of the clinotron oscillator. A 2D single-pixel scanner set-up has been realized so far. In the imaging system, we use a clinotron oscillator that is capable to deliver about 200 mW of CW power in the 300 GHz frequency range. The clinotron wave-guide output is directly connected to a diagonal horn

Session A.1Future Security 2011 Berlin, September 5-7, 201115

which forms the radiation pattern with high radial symmetry. It was designed to provide a paraxial free space mode within the angles of the lens illumination. The output radiation is modulated at a frequency of 0.8 kHz by using a mechanical chopper. The radiation then passes through a 3-dB quasi-optical beam-splitter used to decouple the outgoing and incoming radiation while keeping the Tx and Rx radiation patterns oriented along the same scan axis. On the receiver side, the detector is equipped with a same horn which points normally to the transmitter horn axis into the beam splitter. The optical system consists of a dielectric focusing lens that is used to focus the incident radiation and of a flat mirror which is used to scan the scene. The pla-no-convex focusing lens was made of Teflon, and it has the diameter of 220 mm that equals to 220 wave-lengths at the frequency of 300 GHz. The lens thick-ness is 34 mm. The lens is illuminated by the trans-mitting horn with a -10 dB taper at the edge to pro-vide both optimal gain and maximum Gaussian beam transformation efficiency. The lens was designed, manufactured, and tuned to provide the diffraction-limited beam waist of the spot to be less than 24 mm at 5 m distance. [20] The flat flapping mirror has a rectangular shape of 250 mm x 450 mm. The mirror is made of a silver-plated glass plane that shows a flatness of better than 30 m. The mirror is placed on a PC controlled posi-tioner. The beam formed by the lens and positioned by the mirror is reflected normal to the surface of the inves-tigated object and returns back through the lens. The beam then propagates through the beam splitter reflecting it towards the receiving horn which is con-nected to the detector. The signal from the detector is read by a lock-in amplifier. The synchronizing signal for the amplifier is obtained from an opto-coupler placed on the chopper rim.

Figure 3 Diagram of the single-pixel scanner set-up

Further improvements that can be made to the imag-ing system is that the beam focusing optics can be ac-complished by metal reflectors such as an ellipsoidal offset reflector mounted on a rotation stage. That

would boost the optical efficiency due to a reduction of the absorption loss and a decrease of the reflection loss from the thick Teflon lens. The backscatter loss from the convex surface of the refractive lens could also be eliminated. However, even such a relatively simple set-up as shown in Figure 3 makes it possible to use the devel-oped 300 GHz imaging system for various security and other applications.

4 Future of the clinotron-based imaging systems

The potential of the clinotron oscillators looks very well from the point of view of their possible applica-tions in various THz systems. These oscillators com-bine unbeatable properties, such as compact dimen-sions, a high output power, and a wide frequency tun-ing range. However, for their successful usage in perspective high-resolution 3D imaging systems, some further investigation