Two Way Information Management (TWIM) is a Feasibility Study to assess the added value of receiving information regarding incidents, public safety issues, criminal behaviour etc. from the general public via their smartphones or other mobile devices, in support of operations of Civil Protection Services (Police, Fire Brigade, Ambulance Service etc).
The information collected by TWIM is provided as dedicated information streams each targeted to one specific subject (forest fires, earthquakes, traffic accidents, chemical leaks, meteorological hazards, etc). Each submission is analysed by various analytical methods and tagged with a confidence score allowing the users to filter the messages according to their needs. In return, witnesses are kept informed by the Service Providers of evolutions in the particular incident context they are contributing to.
The objective of this study was to assess the technical feasibility as well as the economic viability of an integrated service that allows the general public act as witnesses to report various types of incidents which are then disseminated to numerous end users.
The concept was based on the popular crowdsourcing movement that has caught traction with the population in the last decade, especially regarding emergency reporting (i.e. Haiti, Katrina etc). The feasibility study combined real time location data using GPS to plot submissions on a map to create a live view. The study also tried to establish the usability of a confidence score associated with each of the images - based on various analytical services. These could then be used to filter and rank the submissions.
As well as the technical aspects, the study looks at the appetite of the general public for this type of application, as well as if the end users (i.e. emergency services) can find any use for this type of tool, and how the business plan would be built around that demand for it.
Users and their needs
The direct users of this service are identified as national or regional authorities responsible for the various emergency services that respond to incidents. Specifically this would be the police, fire and ambulance services. Additionally to that environmental agencies are also imagined to be obvious users of this service allowing witnesses to report flooding or earthquake data would allow the Environmental Agency or the British Geological Society do analyses on the data. Indirect users of TWIM may be the more social or financial institutions for instance insurance companies would be interested to see the data about flood or subsidence damage, at the same time media institutions may well want to report on events that have been reported.
Target users who have also acted as the source for various requirements are: SDIS 31(a fire brigade in Toulouse, France), ACPO (Association of Chief Police Officers, UK) and the Haut-commissariat ? la Protection nationale in Luxemburg (HCPN - the Luxemburg Civil Protection Agency).
The British Transport Police has volunteered to be part of our test campaign and have become another target user for the end product.
The civil protection agencies recognise the need of exploiting the social media and the modern ways of communication in order to access the citizens requests. The witnesses of an event possess valuable information that can improve the civil protection agencies reaction to a call regarding that event. They are also aware of the fact that opening up to new and simpler ways of reporting might result in unmanageable amounts of data and that, therefore, a system that is able to organise (filter) and rate the information before is presented to them is necessary.
Such a system will allow them to attend the relevant calls in a more efficient manner.
Service/ system concept
This is the list of services provided by TWIM:
- Collection of messages from witnesses.
- Online map display: geographical representation of the messages.
- Message filtering service: messages are categorised and presented only to the relevant end user based on their preferences.
- Messages rating service, based on message analysis: each message will be subjected through a variety of analysis services (image, text, location...) and given a confidence score.
- Sharing of TWIM messages between different end users.
- Archive message search: useful for investigations or for future prediction services.
- Communication to witness: one-to-one or one-to-many based on location, time, event, etc. This also closes the loop and allows the witness to be informed of the status of their reports.
Space Added Value
Knowing where an event is happening is the key of the system. Often, the civil protection agencies receive calls with vague information about the location. The satellite positioning will allow the TWIM Submission to have a location stamp that will allow a more accurate positioning.
Since TWIM would be useful in emergency situations, satellite communications are essential. The system needs to account for the possibility that terrestrial communications are down or that the event might happen in a remote location where there is no terrestrial coverage.
As described in the key issues section, there is currently no real way to send location information or images of incidents to a centralised repository that would disseminate and distribute this appropriately.
The benefits to the emergency services are numerous:
- Faster and more detailed understanding of situation using location data and images.
- Better ability to judge response type and assess risk to deployed responders.
- Ability to communicate with message submitter in order to gain more information.
The benefit for the citizen is clear, rapid communication, the "feel good" factor of having potentially helped a situation and the possibility of following on the submitted report and its resolution.
Additionally, TWIM can also help with sharing information between end users which will help coordinating in the case of an incident.
TWIM endeavoured to increase the ability for the various emergency responding agencies to have a better overview of a situation before being on site. This was made possible by sourcing information from the general public via a smartphone application instead of relying on just speech from a telephone call. It must be stated that TWIM was not trying to replace any emergency phone services that exist in any country - just to enhance the information that may be available. Additionally to receiving information from the general public (witnesses), TWIM also aimed at being able to be deployed to on or off duty personnel (trusted witnesses) who would report issues that carry more 'trust'. Each message that was received is analysed by analysis engines (text, image, GIS) to evaluate a confidence score and filter out any redundant or offensive messages.
For the fire services, for example, this service allowed users to view a map of submitted messages identifying clusters of messages. Each message may be accompanied by an image which would give the fire station operator a better indication of what resources needed to be sent. So rather than waiting for the first responders to reach the incident and then calling for backup, the initial first responders could be more numerous or prepared. Similar advantages were be identified for incidents that the police would respond to, or the ambulance service.
By nature TWIM enables information to flow two ways - and this was made possible by the fact that the end users (fire services etc.) are able to write a message to the submitter or an individual message. This could be to ask for more information, or just to thank them and inform them their information is valuable and is being used.
Currently, the only way that civil protection services can receive information from users at a near live time frame is by telephone call. Recently, the British Transport Police have introduced a text message service on the railways. However these messages are limited to text, and for situational awareness, pictures are highly desirable. Furthermore, knowing the exact location of where incidents have been happening or are happening is also hugely beneficial to coordinate an efficient response.
The other issue relates to the crowd sourcing aspect of this approach; how can a system filter reports out that are unrelated / spam and only highlight the most important and pertinent.
TWIM allows users to report something that 'ought to be reported'. Be that a criminal act, accident, incident, natural disaster etc. All of these messages are submitted to the TWIM central repository and distributed to the correct end users automatically. Each message is accompanied by a confidence score calculated by the various analysis services.
The feasibility study has completed with all the major components implemented and tested, including a verification phase and analysis of outcome.
The implementation of the proof of concept was one of the final areas of development, and this including a bespoke mobile app and a website for the end users where the reports will be shown in the different categories and placed on a map. The messages were successfully processed by a prototype of the filter engine for images and text, with some success. During a test campaign at the British Transport Police the timeliness and accuracy of the system were put to the test, and TWIM performed to expectations in both regards. The lessons learnt and future ideas were amalgamated into the final report document along with the roadmap and viability analysis.
The TWIM feasibility study is completed, and work is being done to research into ways of progressing the initial research into something useful for a paying end user.
For information on the results of the feasibility study, reference is made to the final report. A link to the final report can be found at the foot of this webpage.