Blog

Transforming a UAV to an actively connected surveillance vehicle leveraging the use of multisensory data.

An effective assessment of illegal border crossings is planned through the exploitation of active data fusion from various types of sensors, complementing each other’s capabilities. This process uses aerial or ground sensors for devising optimal strategies on reducing errors (type I and II) in airborne detections. To accomplish this, the dominant role of ICT, sensorics, machine learning systems and unmanned aviation technology will assist to overcome challenges such as the deluge of uncertain sensor data and a high resource-consumption cost. The modern border surveillance platforms will have to improve the sensor range as well as the coverage capabilities and expand the mission time, to improve the effectiveness of surveillance along all (remote and nearby) border sections. Towards this goal within BorderUAS project we will fuse and analyse multi-stream data sources, as to produce knowledgeable insight into hidden data patterns for rapid decision-making that would respond to the strict border security requirements. In modern systems the need is to maximise the environment sensing, and bridge the interconnection between the physical world and the computer sensed entities through cutting edge data analysis and context awareness methods, supported by data fusion, feature extraction and perception, while assuring efficient network communication for timely and secure border monitoring and surveillance.

Concept of BorderUAS Sensors & Fusion Integration

BorderUAS sensorics (a technical overview): It will include optical and hyperspectral camera arrays, LTA-UAV-specific synthetic aperture radar (SAR), laser detection and ranging (LADAR), shortwave/longwave infrared (SWIR/LWIR) and acoustic cameras for both, direct as well as indirect target detection, e.g. via vegetation disturbance. The SAR imaging capability will be implemented using a MIMO radar that does not require the platform to be in motion while acquiring SAR images, leveraging the unique characteristics of the Hypersfera’s UAV which can hover above a given target. Besides the imaging capabilities, each antenna of the MIMO radar will be used for moving target identification. All data streams will be processed through a developed data processing, fusion and interpretation layer in order to produce “big” high-dimensional streaming data, appropriately processed, fused and interpreted, that enable effective and efficient detection in difficult terrain. The aim is that BorderUAS solution specifications will drive the service transformation to beyond state of the art technology that will provide ultra-high resolution multi-sensing array payload for the project’s border surveillance vehicle.

Overall, BorderUAS aims to combine an unmanned aerial vehicle with an ultra-high resolution multi-sensor surveillance payload, in order to support border surveillance applications and, specifically, detection of activity in difficult terrain. The primary objectives of BorderUAS sensorics layer include (i) sensing the critical information from the external physical environment, (ii) the sampling of internal system signals, and (iii) obtaining meaningful information from sensor data to perform effective and efficient decision-making.

Gathering border guard authorities’ needs and requirements concerning the most-used migration routes

According to the latest FRONTEX report (2019 Risk Analysis) on illegal crossings of the EU’s external borders, Eastern Mediterranean, Western Mediterranean and Western Balkans are the three major illegal migration routes.

For 2019 the total number of the reported detections of illegal border crossings into the EU was 141.741. A number lower than the year before (2018) however still a significant one.

These routes are characterized by large mountainous and forested environments with difficult terrains causing difficulties in daily monitoring processes conducted by border guard authorities. The report however does not feature any statistics on undetected border crossings.

Do undetected border crossings are more likely to happen in difficult terrains due to surveillance gaps? This seems like a reality since border and law enforcement agencies face critical issues in detecting persons and vehicles illegally crossing EU borders in such regions.

Foliage penetration is one of the most challenging obstacles to overcome in border surveillance especially in the 3 abovementioned routes. Initial work for foliage penetration radar systems dates back to the late 1960 to mid-1970 with poor results due to foliage attenuation that limited the systems to short-to-medium range operation.

In the late 1980s, the development of radars and wideband data links enabling image processing/interpretation on the ground started to provide acceptable and useful detection and characterization of forested regions.

When we talk about radar-based surveillance equipment, a fundamental trade-off between resolution and penetration capability exists: high resolution demands a high centre frequency, but high penetration demands a wavelength long enough to propagate through tree cover.


Source: FRONTEX Migratory Map

The EU funded project BorderUAS (GA no. 883272) introduces an aerial-surveillance approach by combining different types of ultra-high resolution (cm-scale) sensors using technology such as synthetic aperture radar (SAR), laser detection and ranging (LADAR) and shortwave/longwave infrared (SWIR/LWIR).

Through specific pilot cases, system development and assets combination, BorderUAS will test and deploy a combined solution which will provide high coverage, resolution and revisit time with a lower cost than satellites or other aerial surveillance means, such as helicopters, and higher endurance than drones.

As an initial step, the design and layout of the BorderUAS system will ensure first of all the alignment with all operational needs and challenges facing the EU border region. As such the project focuses on an end-user-centred system design that addresses effectively border guard surveillance needs.In order to validate the final outcome in real-life operational environments, BorderUAS organised two workshops with border guard authorities to determine, refine and finalise end-user requirements and trials’ use cases and cross-border scenarios.

The border authorities that participate in the project will continuously contribute to the assessment, refinement and finalization of the use cases, detailed pilot scenarios and requirements during the course of the project, in order to deliver an optimized version of the BorderUAS system dedicated to current extended needs.