foto_scien01a) Targeted interaction with the crowd at the place of the incident, but appropriate in terms of risk exposure. Concerning the Mobile Crowdsourcing Solution, our scientific objective is to engineer mobile applications whose user interface and interaction model support safe and efficient provision of information in emergency situations and result in a positive user experience. This requires understanding and modelling the behaviour of eyewitnesses and first responders in an emergency situation and how their interaction with mobile devices can be supported, so that they can provide precious information to a command centre as soon as possible. In terms of safety, it is essential to provide a very simple and direct way of providing information about the occurrence of an emergency situation when a person is in great danger and allow details to be provided when the degree of danger decreases. As emergency situations cause cognitive load and stress, creating a positive user experience in those situations is a challenge the RESCUER project will address.
foto_scien01b) (Semi-)Automatic fusion, aggregation and analysis of multimedia data. From the point of view of the Data Analysis Solutions, the most frequent data input will consist of user-generated videos/photos taken from multiple points with variable content and image quality. Many tracking and detection algorithms have been proposed in the literature, but the analysis of low quality and moving video/images offers a new challenge, as this kind of multimedia data presents some difficulties that ordinary techniques are not able to resolve successfully. Furthermore, additional information may be combined, which includes text information, open data, and external data from operational forces. Therefore, the scientific objective is to develop an image/video analysis system, based on computer vision techniques enriched by extra scene information, to extract relevant information related to the emergency scenarios of interest. The complexity of this analyser is related to the content variability and the need to reject low quality or inappropriate pieces and combine information from several heterogeneous distributed data sources.
foto_scien01 c) Optimised aggregation of intuitive metaphors for visualisation and manipulation of information. With regard to the Emergency Response Toolkit, our scientific objective is to define the most appropriate mechanisms for visualisation of crowdsourcing information and for coordination of the first responders’ actions in a command centre. This should take into consideration the different phases of an emergency, the decisions to be made in each phase, who should participate in those decisions and the aspects to be considered in each decision. The challenge is to improve situation awareness, while only providing information that is relevant for the decision to be made, thereby avoiding cognitive load.

foto_scien01 d) Customised communication of the incident and its consequences targeted to the audience. Another scientific objective regarding the Emergency Response Toolkit is to support context-aware announcements of emergencies, meaning semi-automated derivation of messages taking into consideration the specific emergency situation, its characteristics, the current emergency phase, and the target-audience (e.g., press, public authority, factory where the incident took place, or the overall industrial area).

foto_scien01 e) Ad-hoc communication between people in the crowd. In many emergency situations, mobile phone based communication (including data transmission) breaks down. The scientific objective related to the Communication Infrastructure is to develop a peer-to-peer communication method based on the built-in WiFi capability of mobile phones in order to provide an ad-hoc communication infrastructure as an alternative to dealing with the unavailability or overload of the usual communication infrastructures.