AI system for Monitoring, Alert and Response for Security in events

AI MARS: Artificial Intelligence system for Monitoring, Alert and Response for Security in events

National Project

The objective of the CIEN AI MARS project is to research various technologies, techniques, tools, methodologies and knowledge aimed at developing technological solutions to support the surveillance and prevention of attacks and other incidents (crowds, riots, etc…) in large concentrations of people (airports, fairs, events, sporting events, parties, shopping malls or similar environments); as well as other elements of interest such as Critical Infrastructures, borders, etc… The AI MARS project facilitates the adoption of technological solutions to provide useful and real-time information to Public and Private Security Forces and Bodies as well as to the managers of large public spaces (shopping centres, sports areas, etc.) to prevent attacks, crowds, riots and other incidents in large concentrations of people and other elements with high security requirements. The social challenge of this project is to improve the security of people by reducing attacks, riots, agglomerations, etc., especially in large concentrations of people, but also applicable to border control or critical infrastructure protection.

Technical Objectives

The technical objectives of the AI MARS project are to investigate a wide range of technologies for application to security and surveillance at large events:
  • Big Data / Machine Learning / Artificial Intelligence and intelligent algorithms
    • Use new state-of-the-art techniques to increase the robustness and accuracy of identifying unique profiles of objects or persons in information received from various sources. These sources may present unstructured, partial, noisy or missing data among other problems that should be addressed on a case-by-case basis
    • Reduce the computational demand for integration algorithms and their latency, by discarding irrelevant or redundant information
    • Define the aspects necessary for the storage of information collected for an individual at different moments in time, aggregated from heterogeneous sources
    • Develop new time series algorithms that are capable of expressing information in different dimensions (for each of the aggregated data sources/sensors)
  • Virtual, Augmented and Mixed Reality
    • To develop techniques that allow visualizing and analyzing the data generated by the project sensors under an immersive visualization in virtual environments and real pseudo.
    • Research on technologies to address the problem of absolute positioning in large spaces
    • Study of the integration of the different markers, including the combination of different technologies, in the equipment carried by the security forces and in other elements such as vehicles, urban furniture, suspicious objects, etc.
  • Real-time image processing
  • Human-Machine Interfaces (HMI)
  • Biometrics (iris, facial, vascular)
    • To reduce the time needed for the calculation of iris identification, so that pre-processing and comparison processes can be carried out in the order of 50ms.
    • Use new state-of-the-art techniques in Deep Learning to increase the robustness and generalization of facial recognition in unrestricted mass environments
    • Design of high performance vascular recognition algorithms based on intelligent systems and high performance architectures
  • Machine vision and recognition of weapons, objects, vehicles and behaviour patterns
    • Develop algorithms for identifying abnormal crowd behavior. Behaviors such as: crowding, falling, jumping fences, stampedes or agitated/stressed masses will be detected
  • Sensorization and the Internet of Things (IoT)
    • Research into Ultra Narrow Band technology in the field of safety, introducing the future concepts of IoT networks
  • Wireless communications: LTE, 4G-5G, WIFI, LPWA…
    • Researching LPWA security-oriented communication technologies
    • Research in hardware solutions for the acquisition of signals in the field from sensors or equipment and their transmission by means of LPWA technology
  • Communications tracking systems
    • Obtain knowledge through research in mobile location systems using WIFI and Bluetooth technology
  • Neutral Operator:MEC, Virtualization, Slicing, mMTC, uRLLC, eMBB, DAS y Nanoceldas
    • Research and Development of high capacity eMMB 5G network.
    • Research and Development of a low latency uRLLC network with edge processing or MEC.
    • Research and Development of a secure network using slicing and virtualization of the 5G network
Project partners
Project collaborators Duration: 2018-2022 The project has been funded through the CDTI‘s CIEN Call for 2018