India-UK Advanced Technology Centre (IU-ATC) of Excellence in Next Generations Networks Systems and Services Phase 2

Lead Research Organisation: Queen Mary University of London
Department Name: Sch of Electronic Eng & Computer Science

Abstract

Abstracts are not currently available in GtR for all funded research. This is normally because the abstract was not required at the time of proposal submission, but may be because it included sensitive information such as personal details.
 
Description 1. IUATC Group 1 WP 1.5
The PI and her team from QMUL was a major research contributor in Work Package 5 of Group 1 (named WP1.5) of the IUATC Phase 2 project. WP1.5 focused on the provision of resilient emergency services during disaster situations. These include events such as earthquakes, landslides and terrorist attacks. The major focus was the provision of an emergency warning dissemination, which was based around an adaptive information dissemination architecture designed to exploit all existing forms of connectivity in order to ensure that users could receive a given piece of information in a manner and format appropriate for their needs. The other partners of the work package were: CEWiT, IIT Delhi and IIT Mandi from India, and Cambridge and University of Ulster from the UK. The QMUL PI was the leader of the WP in UK and CEWIT in India. Concretely, the following tasks were performed within the partners:
Task 1: The development of an integrated disasterrelated information dissemination architecture, which exploits multiple forms of connectivity to ensure resilient and efficient information communications between multiple parties.
Task 2: The development of adaptive (multimodal) emergency warning client applications that can adapt to reflect the needs of their users.
Task 3: The development of adaptive emergency management that can adapt to reflect operational ability and realtime state of the network.
Task 4: The creation of an information model that can be used to capture important characteristics used when describing emergency warnings and users in the system. This is to underpin the integrated system architecture by allowing all components to understand each other.
Task 5: The development of demo scenarios to highlight the general potential of the WP's research to a wider audience.
QMUL was the major contributor in tasks 1, 4 and 5. QMUL also developed warning client applications (task 2) to test the dissemination middleware in the trials in the Lake District along with Cambridge. The outcomes of these tasks are briefly described below:
A. Architecture and Prototype Functionality
After the gathering of system and user requirements and the analysis of requirements, published in the two internal deliverables of the WP1.5, the Early Warning System (EWS) architecture was defined, where the main objectives were:
1. To allow the logically centralised generation of emergency alerts, either manually or automatically, in response to external triggers.
2. To allow the dissemination of emergency alerts to individuals and entities that are in some way affected; This must allow the selected targets to receive the warnings within the constraints stipulated by the disaster's context.
3. To allow the context aware adaptation of the warning messages to reflect the constraints of the end users, the users' device capabilities and the infrastructure availability.
The goal was to design a system that could leverage existing information and communication technologies as well as infrastructure, while ensuring flexibility and extensibility. In particular, the early warning dissemination mechanisms have to be built upon currently available messaging systems, using unicast, broadcast and multicast transmission, as per the disaster, network and user context. It should also enable both centralised and distributed modes of operation, from a functional perspective, for maximum implementation and deployment flexibility. With these considerations, the proposed hierarchical EWS architecture was interconnected by an information dissemination mechanism (e.g. a middleware).
The EWS functions were implemented into four major components: The Warning Generator, the EWS manager, the EW distribution Service and the Disaster Information Service. A prototype implementation was performed. QMUL fully implemented the EWS manager, defined the all the information model used in the message dissemination and also implemented a version of the EW Distribution Service that was used in the demonstration in Surrey in in the trials in The Lake District, which will be described in details later. The functionality of the implemented components is described below.
The EWS Manager has the overall responsibility for generating alerts and ensuring that warnings are received by the target set of users. It also maintains a database of registered users, along with the associated user profiles. Users are allowed to register via SMS or Web. When the EWS Manager receives a warning trigger, it determines the set of target recipients, based on the nature of disaster. These recipients could be explicitly known to the system via registration or, alternatively, be generalised into an anonymous class, to be contacted via broadcast transmission (e.g. cell broadcast). Then requests the Warning Generator to create a suitable message (or a set of messages).
The Warning Generator stores a set of templates for different types of disasters such as floods, landslides etc. In the implementation, the Common Alerting Protocol (CAP) was used. It may return a single, unified message but the EWS Manager may then decide which part(s) are to be distributed to whom and in what format. In the simplest case, the same content is sent all the targets but that may not always be the case. Sometimes, there may be a need for some level of adaptation based on the nature of the communications channel and end device; Images, for example, cannot be shown on a basic phone. Along with the warning, pointers to pull­based services may be provided, in the form or URLs, helpline numbers etc. This information can be used by interested parties to access the corresponding Pull­based Information Service(s). Depending on the target set, the EWS Manager decides upon the appropriate delivery mechanism(s) to be used for warning(s) dissemination. There are 3 delivery mechanisms currently supported by the system. The SMS­based dissemination is available to all users who have provided their mobile numbers at the time of registration. SMS warnings can take two forms: a basic warning that can only contain text information and a more detailed warning which may include text, audio, graphics, links etc. The former essentially caters to the lowest common denominator, i.e. users with basic phones. The second type is targeted towards users who have phones with more advanced features and support installation of application but may not have full­fledged mobile Internet connectivity. The third mechanism is IP­based messaging, based on UDP/TCP, targeting smart devices with wired/wireless Internet connectivity. These diverse distribution mechanisms are enabled via a message dissemination middleware.
The middleware developed by QMUL embodies the above delivery mechanisms in puggable components that can be attached on­demand. The middleware also maintains a repository of user information, detailing the characteristics of individuals, alongside the mechanisms through which they can be contacted. This is exploited whenever a piece of information (such as a warning) needs to be disseminated. The repository maps the target audience of the information dissemination to the users who should receive it (e.g. send to all civilians), before adapting to utilise the optimal mechanisms to reach them. Clearly, different mechanisms will support different methods of transit; For example, a video could not be sent via SMS. The middleware therefore also adapts the contents of the information automatically to reflect the limitations of the communications channel and end device. Through this, the EWS gains the ability to reach a far more diverse group of users.
At the client side, the EWS system is supported by an Information Model to represent User Profiles and the Warning Messages. This is essential for building an open and interoperable set of components. It is formalised in a collection of XML schemas, which all elements of the system possess. The warning messages are compliant with the OASIS Common Alerting Protocol standard, whereas the User Profile model has been developed specifically for the EWS, Cambridge was the major contributor in the User Profile development. The User Profile provides a description of a particular user. It captures any pertinent characteristics that might relate to the network or other related stakeholders, e.g. the host and environment. Although targeted at the emergency domain, it has been designed as a generalised model suitable for many other domains.
B. EWS implementation releases during the project
June 2013 - QMUL fully developed a middleware prototype for operation on both PCs and Android phones. The implementation supported AMQP and SMS dissemination, adapting information to different delivery channels. Any application requiring the consumption of information could be built over this middleware.
September 2013 - QMUL released the first version of the information dissemination middleware, which was used by both Delhi and CEWiT in their implementation work. Further work performed by QMUL on the middleware introduced support for interaction with basic features phones via SMS (i.e. they do not needed the clientside middleware installed locally). QMUL started to add delay tolerant networking support, where messages could be passed from authorised parties (e.g. mountain rescue) to civilians. During this period, CEWiT implemented the user registration and warning generation functions. The same was integrated with the QMUL middleware and tested with IITD client application. QMUL also built an emergency warning Android app that interfaced with the QMUL middleware to alert users about emergency situations. It supported the CAP standard, the information models developed in the WP and also integrated the User Interface (UI) adaptation work developed by Cambridge, which supported font size/colour adaptation, as well as background colour adaptation. A second client app was developed by IIT Delhi, which supported the translation of warnings into the local language (currently Hindi) and also has an option to read out messages in both Hindi and English. This entire prototype was demonstrated in the IUATC general meeting in Surrey (UK) in September 2013.
June 2014 - QMUL released the second version of dissemination middleware, which now supported ad hoc communications between devices. CEWiT has developed an SQL database for storing userrelated information. QMUL also released a new demo application that allowed users to generate SOS messages and send them to emergency responders.
C. Demos and Trials
The first integrated demo was shown at the IUATC General meeting in Surrey (September 2013). This was shown during the presentation for the Group 1 meeting. It involved contributions from all key partners as follows:
1. Delhi showed an emergency warning Android app that could remotely receive warnings and render them based on user requirements (e.g. language, colour-blindness);
2. QMUL showed the dissemination middleware capable to adapting and transporting information objects using a variety of communications technologies this provided communications support for Delhi's app;
3. CEWiT hosted the warning manager and sent the alert to the UK remotely via the QMUL middleware;
4. Cambridge showed their user interface adaptation mechanisms, integrated into both Delhi's and QMUL's GUI implementations
The only real trials performed with the EWS system and apps were performed by QMUL and Cambridge in the Lake District (Coniston) in September and December 2013. This involved the testing of the EWS with ad hoc communications between phones through two apps, one simple SOS App and the Warning App. The use case scenario used for the trail was: A user will go to a specified website that provides full instructions on how to register for the mountain warning service, alongside any further information about themselves (e.g. physical disabilities) and install the Apps. Once the registration has taken place, the user's information will be recorded in the Mountain Emergency Warning System (EWS) for usage during a weather warning or emergency. Example: In the event of an emergency (e.g. heavy snow) and the users are located in a radio dead zone village the following steps will take place:
1. A number of Mountain Rescue volunteers will begin to walk around the village carrying smart-phones with the EWS dissemination App. These smart-phones will be configured to participate in a Store and Forward dissemination using self-configuring wireless networks.
2. As they gain ad hoc connectivity with other devices, they will pass the emergency message (where appropriate) to the user App that also works in a Store and Forward dissemination fashion.
3. If possible, the recipient's device will provide feedback to the EWS dissemination app in the MR volunteer device.
The following trial plan was tested:
1. Turn civilian phone on and place it in the window.
2. Civilian should press SOS button.
3. Responder should position himself outside the building in the car. The responder should then move about in the following ways:
a. The responder should drive past the house (likely won't pick up)
b. The responder should walk past the house (hopefully will pick up)
The trial performed in December 2013 involved placing a civilian phone in the window of the Coniston Mountain Rescue Centre. Cellular communication was completely disabled only the ad hoc communication transmissions were used.
Sending an SOS: This involved clicking the Send SOS button on the civilian's phone, placed in the Window. The weather was light rain. In both cases, when driving past (both fast and slow), the SOS was not received by the rescue volunteer. The device was placed on the dashboard. When walking past the building, the message was received when approximately 28 steps away from the building door.
Sending a Warning: This involved generating a new warning (with a picture) and then sending it to the civilian. First, a test message was sent within the same room. The driving cases both failed: the warning was not received by the client. Once the car was parked directly outside the rescue centre, the two nodes connected and the warning was sent successfully.
(Note that the PI took maternity leave from the period of 25 December 2013 until October 2014).

2. Other work performed by QMUL PI and team:
2.1 The research work in indoor/outdoor localisation
This research work started in the IUATC Phase 1, continued to be developed by the QMUL team during phase 2 with new methods being applied and tested. The outcomes were more research publications and new research paths being followed. The first one, was applied outdoor localisation in radio dead zones in mountainous sites for the localisation of civilians in distress. This proposal was awarded a pilot project (ITaaU) and tested in two real trials in the Lake District. The second was to apply indoor localisation using hybrid wireless technologies to enrich the user interaction with public and commercial buildings. The scenario being researched is a Library where the user is guided by an interactive system with maps to the place or activity of interest. The third is the use of indoor and outdoor localisation with sensing to recognise patterns of behaviour in depression patients, The third is the use of indoor and outdoor localisation with sensing to recognise patterns of behaviour in depression patients, this research also integrates the dissemination middleware developed in the EWS manager by QMUL and findings related to sensing of the pollution sensor research.
2.2 The research work in pollution sensing
This research work started in the IU-ATC phase 1 continued to be developed by QMUL team during IUATC phase 2. The outcomes were novel propositions in data-centric energy efficient adaptive sampling techniques for wireless pollution sensor networks and consequent research publications. Miss Gupta was awarded a PhD degree with this work.
Exploitation Route The prototype of the EWS trialled in the Lake District showed the benefits of such a system to help rescue volunteers. A more developed system could be very useful in the first stage of response of a disaster or large scale emergency, which would have a scale of hours to a week and would involve allowing means of request for help (SOS) and respective search and rescue (SAS) response. The main challenge is to supply integrated and robust communication channels. The information exchange must be resilient, ad hoc if necessary and the actual information recorded and properly tagged. The research developed in the EWS and the ad hoc communication solutions could support the challenges above described, and provide a basic prototype of a usable system.
(Added in 2017) From the work in localisation and the Proof of Concept implemented and trialled by the funding of ITaaU, there are many developments for mountain rescue operations and map integration. The localisation network evolved to much smaller devices and more sensing is being integrated to the localisations nodes. One important evolution is the sensing of people in the mountains in a completely un-obstructive and private way that can gather important information to police and rescue teams in rescue operations or warnings. This work will soon lead to publications and it is a direct result of the IUATC phase two grant that lead to the ITaau grant. (Added in 2018) More developments have been made on the localization network for increased efficiency and low power consumption in wild areas with no infrastructure. Also new research is being performed in integrating radar sensing for human activity detection in outdoors.
Sectors Communities and Social Services/Policy,Environment,Healthcare,Leisure Activities, including Sports, Recreation and Tourism,Other
URL http://www.iu-atc.com/
 
Description Although inside QMUL Library only, the indoor location system developed by QMUL is being used by some students to find more easily items in the library through an interactive map, as the system described in the key findings in section 2.1 (not included yet). The localisation system research evolved to one Bluetooth Low Energy i-beacon and receptors for indoor localisation. This work is being used now together with sensors in a Proof of Concept for helping patients with depression. This PoC is being funded by QMUL via a small grant of the Digital Health initiative. This work is producing results and soon there will be publications. Another research development following the outputs in indoor localisation is a novel indoor localisation for confined spaces using inverse beaconing positioning. The results of this new research are being applied to the monitoring and recognition of activities of daily life (ADL) applied to assist depression patients at home. The results of this research have been partially published and the localisation system is being integrated to a hybrid ADL recognition system. It is expected that this track record will allow the application for a greater project with funding.
First Year Of Impact 2011
Sector Environment,Healthcare
Impact Types Societal
 
Description IT as a Utility Network +
Amount £35,500 (GBP)
Organisation Engineering and Physical Sciences Research Council (EPSRC) 
Sector Public
Country United Kingdom
Start 07/2013 
End 06/2014
 
Description Research work in user and data protection systems 
Organisation BT Group
Department BT Research
Country United Kingdom 
Sector Private 
PI Contribution From the meetings and work with BT during IUATC phases 1 and 2, a collaboration emerged with the security group in BT, QMUL team developed also research work in user and data protection systems. A fully funded QMUL student (QMUL scholarship) received co-supervision from Dr Theo Dimitrakos (BT) and had an internship in BT. The outcomes of this research were: A PhD thesis that articulates the property of specialisation in adaptive software systems and propose a novel methodological framework for the realisation of policy-driven adaptive systems capable of specialisation via adaptive policy transformation. Furthermore, this thesis proposes three distinctive novel protection mechanisms, all three mechanisms exhibit adaptation via specialisation, but each one presenting its own research novelty in its respective field. They are: 1. A Secure Execution Context Enforcement based on Activity Detection; 2. Privacy and Security Requirements Enforcement Framework in Internet-Centric Services; 3. A Context-Aware Multifactor Authentication Scheme Based On Dynamic Pin. The other outcomes related to this work are three research publications, and three patents (which belongs to BT) along with the awarded PhD degree (Mr Yair Diaz-Tellez) in QMUL.
Collaborator Contribution Co-supervision of the research, payment of internship to student, payment of travel expenses for conferences.
Impact Three research publications, and three patents (which belongs to BT) along with the awarded PhD degree (Mr Yair Diaz-Tellez) in QMUL.
Start Year 2011
 
Description Visit and workshop to the Coniston Mountain Rescue Team 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach Local
Primary Audience Public/other audiences
Results and Impact Cambridge and QMUL have visited the Coniston Mountain Rescue Team. This involved meeting and interviewing various key people involved in the region's emergency services, thereby allowing the wider community to learn of the IUATC and ITaaU research projects. It also involved hosting a local workshop attended by people from multiple emergency management teams, once again, allowing wider dissemination of QMUL and Cambridge research results. These included demonstration of existing software. The trip also involved establishing the arrangements for future trial deployments.
Year(s) Of Engagement Activity 2013