Autonomous security drones
Transforming commercial UAVs into fully independent security systems through proprietary optical docking and dynamic spatial routing.
Overview & challenge
Developed as the flagship autonomous security solution for skillos.systems, the platform enables true drone-in-a-box capabilities. It can automatically launch from a scheduled patrol or external alarm trigger, navigate a complex perimeter, and return to its exact starting point to land on a wireless charging pad without a human pilot.
Complete autonomy depends on the last mile: landing. Standard GPS systems suffer from daily satellite drift, so a drone returning to base can miss its charging pad by several meters. Automated wireless charging requires repeatable landing accuracy inside a strict 100 mm tolerance, which out-of-the-box commercial drones and standard software cannot reliably achieve.
The solution is built for enterprise teams managing critical infrastructure, industrial zones, and high-security facilities that need reliable 24/7 aerial surveillance and rapid response without a full-time team of human drone operators.
Proprietary know-how
SkillOS combines optical docking, relative routing, and carefully controlled hardware-limit bypasses to deliver repeatable autonomous patrols.
A hybrid computer-vision landing system uses onboard cameras to lock onto a specialized visual marker. Advanced control loops compensate for crosswinds and inertial drift, actively guiding descent until the drone reaches the charging contacts.
The Relative Offset Engine avoids brittle global coordinates by mapping each patrol route relative to the home docking station moments before takeoff, keeping routes aligned with the physical site despite GPS drift.
The architecture manages OEM obstacle-avoidance restrictions during the final descent, allowing the aircraft to complete a smooth autonomous touchdown on third-party charging hardware.
Technical leadership
As Technical Lead and Chief Systems Architect for the skillos.systems drone division, Oleksii Kyslenko led the autonomous architecture from concept to deployment. He designed the mathematical models for inertial dampening and optical tracking, then guided the development team in transforming off-the-shelf drones into self-sustaining security agents.
For a drone to be a true security asset, it must operate without human intervention. We engineered the mathematical and optical intelligence required for a drone to wake up, secure a complex perimeter, and reliably dock itself back to power, day after day.