Final Report Summary - HANDHOLD (HANDHold - HANDHeld OLfactory Detector)
Executive Summary:
For decades, detection (or sniffer) dogs have been used to help humans detect a wide range of substances of interest. Indeed, explosion detection dogs were used as far back as the Second World War to help soldiers find mines that would otherwise have proven extremely difficult to locate. Detection dogs are noted for their incredible olfactory skills, which far surpass those of any human – a dog’s limit of detection can be measured in parts per billion. In recent years they have helped authorities detect invasive species, bed bugs, low blood sugar in diabetics and even cancers, to name but a few.
Developments in sensor technology and low-power embedded systems have led to the creation of other means of detection that do not suffer from some of the limitations presented by the use of detection dogs. In 2010, in its calls for proposals, the European Commission (EC) declared a need for a ‘mechanised dog’, capable of detecting a wide variety of potentially illicit substances quickly and reliably. In response, a consortium established the HANDHeld OLfactory Detector (HANDHOLD) project, with the goal of producing and developing technologies to effectively complement the role of detection dogs.
The HANDHOLD project was funded under the security theme of the European Union (EU) Seventh Framework Programme (FP7) and is coordinated by The Centre for Secure Information Technologies (CSIT), a €40 million innovation and knowledge centre based at Queen’s University Belfast (QUB). It is the largest academic research centre in cyber security and is recognised by the UK Research Council and the UK Government Communications Headquarters as an Academic Centre of Excellence in Cyber Security Research. Led by Dr Charles Gillan, HANDHOLD ran for a total of 42 months.
The core aim of the project was to create a cost-effective and portable means of detection that will assist border agents and civil law enforcement by complementing the role performed by detection dogs. “The Handhold platform can be used as a first indication of a ‘target find’. If necessary, the dog can be brought in later to help locate the specific substance,”. “For example, for passengers walking through a Customs Hall at an airport, it is unlikely that a dog is needed, as an officer can simply speak to the passenger to further the investigation. On the other hand, when searching in a building or dockside, a dog may subsequently be deployed to locate the substance.”
The consortium also includes nine partners from five member states of the EU, which enhances the effectiveness of the project, imbuing it with a diversity of opinion and thereby increasing the scope of its application. “Diversity exposed us to the different types of end use challenges,” Gillan points out. “For example, in the UK and Ireland, -20 °C is typically the operational equipment limit for outside operations in winter, although it is seldom reached. On the other hand, on the eastern European border, the temperature can go well below that in January. The presence of partners who experience this weather kept it at the forefront of our minds.”
Project Context and Objectives:
While the dog will remain a central part of the detection process at border crossing and airports, sensor technology and low power embedded system computing are improving to the extent that the time is right to develop substantially improved mobile detection devices that can complement the role played by dogs. Moreover, these detection devices can be networked together to provide enhanced detection facilities and also to facilitate easier management and field deployment of the platforms themselves.
The HANDHOLD consortium consisted of nine partners who brought complementary expertise in all the fields needed to develop a mobile network based, low power CBRNE sensor system. The consortium included the Irish Customs Authority and the consortium has an attached user group of representatives from law enforcement from around Europe.
Over a lifetime of 48 months, the project delivered a working prototype system in two phases, the initial version being completed 24 months into the project lifetime.
The HANDHOLD platform incorporates novel airflow mechanisms to ensure that suitable amounts of target molecules are distributed around up to four sensors equipped at the same time on the chassis. In addition advanced software engineering techniques enable the concurrent analysis of various spectra from the sensors followed by reporting in real-time to the end user.
Four new sensors were developed for the HANDHOLD platform by the project.
An advanced geographical mapping and planning tool formed a layer of software that makes the capability of the HANDHOLD platform available to mission commanders. This layer, developed by partner INOV, is of particular interest to the Portuguese Judiciary Police who significantly influenced its development.
Project Results:
Two different departments at Queen's University Belfast participated in Handhold. The Institute for Electronics, Communications and Information Technology (ECIT) co-ordinated the project and brought its strengths in computer architectures and embedded systems implementation to WP4 of the project. The Institute for Global Food Security (IGFS) participated in WP6 contributing to the development of new electrode chemistry for the bio-sensor and also in evaluating the PalmSens EMStat as an example of independent third party technology that
could be integrated onto the Handhold platform.
ECIT has developed considerable new opportunities from participation in Handhold. These centre on the low power and embedded computing field and on the data analytics field. In addition to providing command and control for the electronics, the computing on the Handhold platform is a data analytics environment designed to process the spectra that are provided by each of the sensors. During the lifecycle of Handhold
The term Big Data became well known in the computing field and associated to that the role of data analytics. The Handhold platform is an example of power constrained data analytics, a niche situation within the field. ECIT brought this knowledge forward into new consortium proposals in FP7 and H2020, mainly in the ICT and FET Programmes. Nanostreams is one such consortium and
from a technical computing perspective NanoStreams is the closest to the work conducted in the Handhold project. NanoStreams is building an optimised stack, hardware and software, to address the unique challenges of processing real-time streams of data in a scalable, latency sensitive and also energy efficient way. In common with Handhold, the NanoStreams software is executed on an ARM based computer.
Whereas in Handhold the data streams are spectra, in NanoStreams the data is more general. One use case is that of streams of medical data, gathered from sensors attached to a patient. ICU monitoring and alarm systems lack the ability to recognize physiological syndromes such as Acute Lung Injury and its more severe form acute respiratory distress syndrome. NanoStreams algorithms aim to overcome the limitations by employing real-time analytical capability. Thus time dependent streams of tidal volumes and airway pressure readings provide the raw data. Mathematical kernels for trend analysis and thresholding are applied to the data over varying time scales. NanoStreams spawned investigations of new treatment options in respiratory physiology in ICUs. We are engaged with InvestNI to gain funding to create a start-up company building this further. We have just begun a trial of the research prototype ICU analytics application, known as VILI, in a hospital in Belfast; publications will not be prepared until after this is completed in October 2016.
Analytics Engines investigated clustering and segmentation algorithms for data analytics running on an ARM based embedded platform. This experience combined with the software generated to implement these algorithms can be translated to other areas of data analytics including
• bioinformatics
• customer segmentation.
In addition, AEL gained new knowledge working with morphology algorithms for image processing which they have further disseminated in an advisory role to partners working on other research projects.
KIT developed two different type of chemical sensor in the project each using different properties of molecules. One stand of work used Mid-Infrared spectroscopy (MIR) while the other used Fluoresence Quenching (FQ) spectroscopy.
The MIR sensor has been proved to have a detection capability of up to 1 ppm in trials of the Handhold platform. Being based on analysis of spectra, it is a versatile detection system and can be re-targeted to many substances by changing the wavelength filters. The prototype was tested for the detection of tobacco and cocaine.
Potential Impact:
HANDHOLD designed a platform specifically to meet the needs of Customs officers and civil law enforcement officers. It is designed so that other vendors of sensors could, if they wished, integrate into the HANDHOLD system. Teh system allows for the operation of several sensors at the same time, covering a range of CBRNE targets.
Since the project began the need for CBRNE detection at European borders has increased. The immigration crisis in Southern Europe is being exploited criminal gangs to also send narcotics and other illicit material to Europe. The HANDHOLD system is today only a prototype but it is capable of being expanded, with further investment. The primary socio-economic impact that would follow from a commercial product based on HANDHOLD is a better working environment for Customs Officers and their dogs.
List of Websites:
http://www.handhold.eu
For decades, detection (or sniffer) dogs have been used to help humans detect a wide range of substances of interest. Indeed, explosion detection dogs were used as far back as the Second World War to help soldiers find mines that would otherwise have proven extremely difficult to locate. Detection dogs are noted for their incredible olfactory skills, which far surpass those of any human – a dog’s limit of detection can be measured in parts per billion. In recent years they have helped authorities detect invasive species, bed bugs, low blood sugar in diabetics and even cancers, to name but a few.
Developments in sensor technology and low-power embedded systems have led to the creation of other means of detection that do not suffer from some of the limitations presented by the use of detection dogs. In 2010, in its calls for proposals, the European Commission (EC) declared a need for a ‘mechanised dog’, capable of detecting a wide variety of potentially illicit substances quickly and reliably. In response, a consortium established the HANDHeld OLfactory Detector (HANDHOLD) project, with the goal of producing and developing technologies to effectively complement the role of detection dogs.
The HANDHOLD project was funded under the security theme of the European Union (EU) Seventh Framework Programme (FP7) and is coordinated by The Centre for Secure Information Technologies (CSIT), a €40 million innovation and knowledge centre based at Queen’s University Belfast (QUB). It is the largest academic research centre in cyber security and is recognised by the UK Research Council and the UK Government Communications Headquarters as an Academic Centre of Excellence in Cyber Security Research. Led by Dr Charles Gillan, HANDHOLD ran for a total of 42 months.
The core aim of the project was to create a cost-effective and portable means of detection that will assist border agents and civil law enforcement by complementing the role performed by detection dogs. “The Handhold platform can be used as a first indication of a ‘target find’. If necessary, the dog can be brought in later to help locate the specific substance,”. “For example, for passengers walking through a Customs Hall at an airport, it is unlikely that a dog is needed, as an officer can simply speak to the passenger to further the investigation. On the other hand, when searching in a building or dockside, a dog may subsequently be deployed to locate the substance.”
The consortium also includes nine partners from five member states of the EU, which enhances the effectiveness of the project, imbuing it with a diversity of opinion and thereby increasing the scope of its application. “Diversity exposed us to the different types of end use challenges,” Gillan points out. “For example, in the UK and Ireland, -20 °C is typically the operational equipment limit for outside operations in winter, although it is seldom reached. On the other hand, on the eastern European border, the temperature can go well below that in January. The presence of partners who experience this weather kept it at the forefront of our minds.”
Project Context and Objectives:
While the dog will remain a central part of the detection process at border crossing and airports, sensor technology and low power embedded system computing are improving to the extent that the time is right to develop substantially improved mobile detection devices that can complement the role played by dogs. Moreover, these detection devices can be networked together to provide enhanced detection facilities and also to facilitate easier management and field deployment of the platforms themselves.
The HANDHOLD consortium consisted of nine partners who brought complementary expertise in all the fields needed to develop a mobile network based, low power CBRNE sensor system. The consortium included the Irish Customs Authority and the consortium has an attached user group of representatives from law enforcement from around Europe.
Over a lifetime of 48 months, the project delivered a working prototype system in two phases, the initial version being completed 24 months into the project lifetime.
The HANDHOLD platform incorporates novel airflow mechanisms to ensure that suitable amounts of target molecules are distributed around up to four sensors equipped at the same time on the chassis. In addition advanced software engineering techniques enable the concurrent analysis of various spectra from the sensors followed by reporting in real-time to the end user.
Four new sensors were developed for the HANDHOLD platform by the project.
An advanced geographical mapping and planning tool formed a layer of software that makes the capability of the HANDHOLD platform available to mission commanders. This layer, developed by partner INOV, is of particular interest to the Portuguese Judiciary Police who significantly influenced its development.
Project Results:
Two different departments at Queen's University Belfast participated in Handhold. The Institute for Electronics, Communications and Information Technology (ECIT) co-ordinated the project and brought its strengths in computer architectures and embedded systems implementation to WP4 of the project. The Institute for Global Food Security (IGFS) participated in WP6 contributing to the development of new electrode chemistry for the bio-sensor and also in evaluating the PalmSens EMStat as an example of independent third party technology that
could be integrated onto the Handhold platform.
ECIT has developed considerable new opportunities from participation in Handhold. These centre on the low power and embedded computing field and on the data analytics field. In addition to providing command and control for the electronics, the computing on the Handhold platform is a data analytics environment designed to process the spectra that are provided by each of the sensors. During the lifecycle of Handhold
The term Big Data became well known in the computing field and associated to that the role of data analytics. The Handhold platform is an example of power constrained data analytics, a niche situation within the field. ECIT brought this knowledge forward into new consortium proposals in FP7 and H2020, mainly in the ICT and FET Programmes. Nanostreams is one such consortium and
from a technical computing perspective NanoStreams is the closest to the work conducted in the Handhold project. NanoStreams is building an optimised stack, hardware and software, to address the unique challenges of processing real-time streams of data in a scalable, latency sensitive and also energy efficient way. In common with Handhold, the NanoStreams software is executed on an ARM based computer.
Whereas in Handhold the data streams are spectra, in NanoStreams the data is more general. One use case is that of streams of medical data, gathered from sensors attached to a patient. ICU monitoring and alarm systems lack the ability to recognize physiological syndromes such as Acute Lung Injury and its more severe form acute respiratory distress syndrome. NanoStreams algorithms aim to overcome the limitations by employing real-time analytical capability. Thus time dependent streams of tidal volumes and airway pressure readings provide the raw data. Mathematical kernels for trend analysis and thresholding are applied to the data over varying time scales. NanoStreams spawned investigations of new treatment options in respiratory physiology in ICUs. We are engaged with InvestNI to gain funding to create a start-up company building this further. We have just begun a trial of the research prototype ICU analytics application, known as VILI, in a hospital in Belfast; publications will not be prepared until after this is completed in October 2016.
Analytics Engines investigated clustering and segmentation algorithms for data analytics running on an ARM based embedded platform. This experience combined with the software generated to implement these algorithms can be translated to other areas of data analytics including
• bioinformatics
• customer segmentation.
In addition, AEL gained new knowledge working with morphology algorithms for image processing which they have further disseminated in an advisory role to partners working on other research projects.
KIT developed two different type of chemical sensor in the project each using different properties of molecules. One stand of work used Mid-Infrared spectroscopy (MIR) while the other used Fluoresence Quenching (FQ) spectroscopy.
The MIR sensor has been proved to have a detection capability of up to 1 ppm in trials of the Handhold platform. Being based on analysis of spectra, it is a versatile detection system and can be re-targeted to many substances by changing the wavelength filters. The prototype was tested for the detection of tobacco and cocaine.
Potential Impact:
HANDHOLD designed a platform specifically to meet the needs of Customs officers and civil law enforcement officers. It is designed so that other vendors of sensors could, if they wished, integrate into the HANDHOLD system. Teh system allows for the operation of several sensors at the same time, covering a range of CBRNE targets.
Since the project began the need for CBRNE detection at European borders has increased. The immigration crisis in Southern Europe is being exploited criminal gangs to also send narcotics and other illicit material to Europe. The HANDHOLD system is today only a prototype but it is capable of being expanded, with further investment. The primary socio-economic impact that would follow from a commercial product based on HANDHOLD is a better working environment for Customs Officers and their dogs.
List of Websites:
http://www.handhold.eu
