REACTION

REal-time ArtifiCial InTellIgence for BOrders Surveillance via RPAS data aNalytics to support Law Enforcement Agencies

During the past decade, Greece and Cyprus, being both part of the external EU borders, have been called to address major challenges associated with the socio-political situation in their surrounding territories and adjacent countries. In particular, irregular migration flows exert significant pressure on the relevant authorities and agencies that operate at respective border territories. Moreover, armed conflicts, changes in climate conditions, and unpredictable factors occurring at the EU external borders, have increased the number of reported transnational crimes including irregular migration. Smuggling, human trafficking, and irregular migration activities are major concerns for South Eastern EU Borders particularly, as monitoring the routes used is hindered by vast sea and mountainous, densely forested areas and rough lands. Hence, early reaction and identification of relevant events may significantly increase the capacity of the border management agencies and provide adequate reaction time which is crucial and may be reflected to human lives.

Towards addressing these challenges, REACTION aims to integrate, validate, and demonstrate a fully functional, next generation, holistic border surveillance and awareness platform, providing pre-frontier situational awareness beyond remote areas as an effective mean for early identification of critical situations. REACTION will also be integrated and interconnected with the already installed information platforms in Reception & Identification Centers (RICs) and Fylakio (Evros) for proactive actions. Based on the outcomes of three successful EU-funded projects (CERETAB, AIDERS, and ROBORDER), a variety of autonomous vehicles (UxV) and relevant operational equipment will be utilized using a simplified commanding scheme to monitor remote EU border territories. Current AI advancements for detection, recognition, and classification will complement the UxV operations towards a fully automated early detection scheme that relies on improved multimodal fusion models. In addition, innovative solutions will be incorporated to facilitate the process of collecting information for hazardous events where migrants are involved. An enhanced Command and Control platform will be configured to monitor all processes in real and almost real-time. REACTION will exploit application-specific algorithms and a novel mapping platform to harness the large volume of data that Law Enforcement Agencies (LEAs) collect through heterogeneous sensors (including visual, thermal, and multispectral cameras, LIDAR, et al.) and convert that data into useful information for enhancing borders surveillance and improving emergency response. The collected data from the various data sources (legacy E/O, radar, and UxVs) will be integrated utilizing the (enhanced) Common Information Sharing Environment (e-CISE) developed in project ANDROMEDA (GA: 833881) and EUROSUR into sophisticated C2 systems allowing the land and coast guard authorities to gather all the additional data that could be relevant to their operations. Early warning, crisis classification and risk/threat assessment will deliver an improved and automated intelligent picture to local, regional, and national command centers. Towards delivering a TRL-8 fully functional system, real operational scenarios will be defined considering also previous projects’ scenarios, lessons learned, and gaps identified, and will be utilized for execution under real field conditions within actual border territories.

M4D has the role of the Technical Manager, working closely with partners to ensure compliance with quality assurance and work plans, monitor the achievement of project objectives, and ensure the proper execution of simulation exercises and evaluation tests. M4D also leads one WP about enhanced operational capabilities and services. M4D will deploy advanced computer vision and deep learning algorithms to analyze video streams from various sensors. It will also provide predictive analytics by adapting existing models to classify crises and illegal events, improving situational awareness through intelligent data enrichment and threat assessment techniques. Lastly, M4D is responsible for defining detailed architecture and technical requirements, ensuring smooth interaction between software components, and adopting a modular approach for flexible integration.