Autonomous Drone-Based Power Infrastructure Inspection

Authors Details

Antonis Savva, Yiannis Grigoriou, Panayiotis Kolios, Christos Panayiotou

Research Unit

KIOS Research and Innovation Centre of Excellence

Description

Autonomous inspection of power networks using Unmanned Aerial Vehicles (UAVs) has gained significant attention due to the rapid advances in embedded devices and UAV technology. In this context, UAVs equipped with high-end onboard processing units and camera payloads, are dispatched across the power network for acquiring high-quality data safely and fast. This task is particularly challenging especially in cases where the location of infrastructure components, i.e. poles, is unknown. In this work, we capitalize on breakthroughs on Jetson devices to develop a vision-based AI toolkit, which can process in real-time vision sensory input from the UAV’s camera payload and detect poles whose location is unknown. Detection output is integrated with the flight controller for aligning the UAV directly above the pole marking its correct location. The proposed approach has been successfully applied to autonomously inspect ~3.5km of the medium voltage network in an unseen region

TRACE: KIOS Tracking, Rescue, and Aerial Coordination Ecosystem

Authors Details

Michalis Demetriou, Christos Panayiotou, Panayiotis Kolios

Research Unit

KIOS Research and Innovation Centre of Excellence

Description

TRACE is a multi-agent AI toolkit designed to equip incident commanders and first responders with an advanced suite of tools for managing emergency response operations. By collecting, analyzing, and visualizing real-time data from UAV-mounted sensors, TRACE enables the creation of knowledge maps that support the development of effective, evidence-based response strategies. As an IoT-powered solution, TRACE offers both standalone tools for intelligent UAV-based data collection and processing, as well as advanced ground control applications for enhanced data analytics and visualization. The toolkit includes robust command-and-control capabilities for managing UAV teams across a range of missions, such as aerial data collection, search and rescue, and area monitoring. Additionally, the integrated visualization tools provide enhanced situational awareness, enabling users to remain focused on mission-critical tasks while making more informed, real-time decisions.

FixCyprus: Leveraging Crowdsourced Smartphone Data for Road Safety Hazard Management

Authors Details

Andreas Georgiou, Christos Laoudias, Aristotelis Savva, Christos Panayiotou

Research Unit

KIOS Research and Innovation Center of Excellence

Description

Our work focuses on FixCyprus a crowdsourcing platform developed to address road safety hazards in Cyprus. The primary problem it tackles is the inefficiency and high costs of traditional road infrastructure monitoring methods, such as field inspections or sensor-equipped vehicles. FixCyprus enables citizens to report infrastructure defects (like potholes, damaged lighting, or blocked drainage) by submitting geolocated images through a mobile application. These reports are automatically forwarded to the relevant district’s Public Works Department based on location, where they are reviewed and assigned to the appropriate public authority for resolution. The platform integrates multiple technical components, including a user-friendly mobile interface for citizen submissions, back-end systems for secure data storage and management, and dedicated portals for public authorities to manage and respond to reports. From a business perspective, FixCyprus offers a cost-effective solution for governments, reducing the need for costly field surveys and facilitating more efficient resource allocation for infrastructure maintenance. It also fosters civic engagement by involving citizens in public infrastructure upkeep.

PAVEment SCANning with EGNSS technology for accurate assessment (PAVE-SCAN)

Authors Details

Symeon Christodoulou

Research Unit

EUPALINOS Lab, Dept. of Civil and Environmental Engineering

Description

PAVE-SCAN (1) aims for the development to market (TRL8-9) of an EGNSS-based integrated low-cost sensor technologies and artificial-intelligence-driven open-architecture software solution (machine learning (ML) and machine vision (MV)), for the detection, classification, and georeferencing of roadway pavement surface anomalies and for the low-cost assessment of roadway pavements using participatory sensing; (2) is based on past scientific and applied knowledge of the consortium’s core technical partners, that has, to date, yielded a field-tested prototype of the proposed solution (TRL 6-7).