EFFECTOR provides both interoperability for and intelligent analysis of maritime surveillance data, with the end goal of faster, more efficient and ultimately “seamless” cooperation and decision-making by border authorities.
In the project’s own words: “EFFECTOR aims to enhance maritime surveillance, improve decisions support, and foster collaboration of maritime stakeholders by implementing an Interoperability Framework and associated Data Fusion and Analytics services for Maritime Surveillance and Border Security that will allow faster detection of new events, better informed decision making, achievement of a joint understanding and undertaking of a situation across borders, allowing seamless cooperation between operating authorities and on-site intervention forces ensuring that all existing privacy and data protection rules are fully respected.”
The EFFECTOR System Architecture (infographic) is a distributed system that — in the tested version — consisted of six modules:
1) The CISE (Common Information Sharing Environment, ed.) Adapter that “transposed / adapted different data models / services from and to CISE model”;
2) The Input and Output layer to harmonize “the data incoming, ingest and also export the data inside and outside the system (to the CISE Network and possibly EUROSUR)”. EUROSUR is the European Border Surveillance system, “a framework for information exchange and cooperation between Member States and Frontex to improve situational awareness and increase reaction capability at the external borders”;
3) The Data Lake that “stored the data efficiently for the different kind of usages”;
4) The Analytics and Ontology that were “supplied directly by the Data Lake and works on data”;
5) The Data Management layer that “provided security and rules inside the EFFECTOR system and data Lake storage and accessibility”;
6) an autonomous object detection and tracking module, carried on-board a UAV — equipped with “a pair of daylight and thermal cameras providing a video stream for the UAV’s pilot” and, thanks to an “on-board AI module”, capable of detecting objects, tracking the motion of detected objects, provide “timely information” (live-streamed), being rapidly deployed, interoperable, self-sufficient and cost effective, the project claims.
Details about the “autonomous object detection algorithms” developed and the use of Big Data can be found in the project’s Publications.
Writes EFFECTOR’s Cordis page that “The project will leverage on the developments, results and experience from current and previous research projects”.
These include EUCISE202o, MARISA, RANGER (which will include “innovative Radar technologies with novel technological solutions for early warning, in view of delivering a surveillance platform offering detection, recognition, identification and tracking of suspicious vessels, capabilities exceeding current radar systems. It will be a platform, consisting of 2 radar technologies, a novel Over-The-Horizon Radar combined with a Multiple Input Multiple Output one implemented exploiting the latest photonics advancements, and an Early Warning System exploiting deep and adaptable machine learning schemes able to Automatically detect radar Targets”), PERSEUS (“a large scale demonstration of a EU Maritime surveillance System of Systems”, ended in 2015), and BLUEMASSMED (ended in 2012, but still interesting as it makes us realize how objectives and promises back then are similar to today’s).
EFFECTOR will also “exploit the e-CISE Data Model of ANDROMEDA”, building on “lessons learnt” from that project.
In 2021 EFFECTOR, COMPASS2020 and ANDROMEDA joined efforts.
Lastly, as a project that is part of the “BES Cluster”, “synergies” have been developed between EFFECTOR and other EU-funded projects, namely METICOS, PERCEPTIONS, MIRROR, TRESSPASS, BORDERUAS, NESTOR, PROMENADE and again COMPASS2020 and ANDROMEDA.
Project outputs were “Tested, validated and demonstrated in real operational scenarios” in
1) Greece — for the “Detection and apprehension of irregular migrants & SAR Operation” and the “Detection of suspicious activity”; specifically, 1a) “The first scenario detected and apprehended irregular migrants, concluding to a Search & Rescue operation in Evros River Delta. Greek authorities analyzed the high possibility of mass migration in the nearby cities and with the help of the EFFECTOR system, migrants’ departure details were estimated.”
1b) “The second scenario focused on the detection of vessels’ suspicious activity between Samothraki Island and Alexandroupolis where only fishing vessels and passenger ships usually occur. A notification for collision avoidance alerts and a rendezvous was automatically identified by the EFFECTOR system”;
2) Portugal, where “The objectives of the trial were accomplished by the enhancement of the prediction, survey capability and the faster reaction to incidents that are related to illegal activities”; and
3) France.
Details for each trial are publicly provided in a press release. Frontex is listed among the participants.
As a result of these trials, the project concludes:
“The impact observed in end-users’ operational capabilities was significant both at cross-border level and cross-sector by sharing information with other fellow member states or sectorial entities respectively. Furthermore, the Artificial Intelligence utilized by EFFECTOR, significantly supports decision-making, and reduces the resources that need to be assigned in order to obtain an enhanced situational picture for maritime surveillance.”
And yet, we might need even more surveillance, the Consortium argues: “The project’s results provide a robust baseline to build on, either by further developing the EFFECTOR system infrastructure or further enhancing its individual functionalities and in this way the overall capacity of the system or even bring inspiration for future innovative relevant projects”.
Actual use cases are advertised in the following way in a brochure: “Reduce the death toll at sea via Search and Rescue, boost maritime surveillance for illicit activity detection, assist in early warning and monitoring of environmental maritime conditions, irregular migration, illegal fishing etc.”
Frontex also concretely benefited from EFFECTOR outputs. Wrote the 2022 edition of the Frontex report on its research activities that the agency’s “engagement materialised in concrete project results such as with EFFECTOR, an outcome which now brings tested concepts and tools to Frontex operations”.
More specifically, “The project objectives and outcomes translated to Frontex systems. Collaboration was established between the Effector project and the relevant Frontex entity responsible for operational data management resulting in the development of a software adaptor between the Frontex Joint Operations Reporting Application and the CISE7 network. With the use of this adaptor during one of the EFFECTOR’s trial campaigns Frontex successfully performed information exchange with partners from outside EBCG community. The software adaptor was handed over to Frontex for future use. Preparations are ongoing to integrate the developed adaptor in Frontex systems.”
How many different interoperable systems for increased “situational awareness” and border surveillance have been developed through EU funding over the years, and are they interoperable among themselves? Interestingly, the project starts from the opposite premise that “new and novel surveillance concepts are still underutilized” (as well as “data sources”).
Also, “decision support” is assumed to be better because of “faster detection of new maritime events”. But does faster decision-making (based on automated recognition of events) necessarily equal better decision-making?
The “data fusion and analytics layer” of the EFFECTOR system promises “object detection”, “situation awareness” (which means recognising “abnormal behaviour”), and “impact assessment”. The main idea is to pool data and make them interoperable from a wide range of sources: vessels, radars, environmental sensors, CCTV and other surveillance systems (coastal and offshore), radio, wireless and satellite communications, video streams and data from UAVs, but also “participatory sensing platforms” (which means “Images, videos and data from citizens (mobile phones app), social media data”) — all of this with a further “possible integration with EUROSUR services”.
Can this be done while respecting fundamental rights? What “artificial intelligence” is adopted, here? Technical specifications of the adopted AI would be key to judge, especially concerning its predictions/estimations/decision support functions. Notions around the “early warning” indicators of irregular migration according to the system would also be an important addition to public knowledge. And is, for example, the assessment automated? It is basically impossible to answer such questions, as — even two years after project completion — only a handful of deliverables is available on either Cordis or the project’s own website. Also, none of them is dedicated to understanding the ethical and societal aspects of EFFECTOR outputs.
