Radar-based drone detection for enterprise sites
Ozense is developing a cost-efficient, locally operating system that detects and tracks small non-cooperative drones. An edge unit on site is intended to separate real target tracks from clutter and assess them conservatively; for relevant events, optical verification via existing PTZ/VMS equipment is planned.
The project is in research and development. It is a proof of concept in the pre-founding stage – not a production-ready product, and not a counter-drone effector.
Cameras need light. RF receivers need a radio link. Radar needs neither.
What is radar-based drone detection?
Radar-based drone detection identifies aircraft by the radar waves they reflect – not by their radio traffic and not from a camera image. An FMCW radar measures the range and radial velocity of every reflecting object, which means it can also pick up drones flying autonomously with no radio link.
The hard part is not receiving the echo but answering what is reflecting: a drone, a bird, a moving branch, or a processing artefact. Making that distinction under real interference conditions and on low-cost hardware is precisely the Ozense research question.
The approach
Detect and track
A 77 GHz FMCW radar picks up moving airborne candidates independently of lighting and radio link. Individual detections are linked over time into tracks – only temporal stability turns an echo into a credible target candidate.
- 77 GHz FMCW
- 3TX/4RX MIMO
- Independent of visible light
Discard clutter and artefacts
Before any semantic statement comes the question of track validity: is there a stable physical target, or only moving environmental clutter and processing artefacts? This separation is the most effective structural lever against false alarms and is therefore kept as its own stage.
- Adaptive clutter filter
- Track validity as a separate stage
- Ghost-target handling
Assess conservatively
A valid target track is assessed as a drone, a bird or other airborne object, or Unknown. Unknown is a deliberate system decision: where evidence is thin, no class is forced. Ozense makes uncertainty visible instead of presenting an ambiguous result as confirmed.
- Micro-Doppler & track features
- Drone / bird or other airborne object / Unknown
- No forced class
Verify and hand over
Target architecture: on a relevant candidate the system slews an existing PTZ camera, where visibility and target geometry allow, and hands the event to a control room or VMS. PTZ cueing and end-to-end integration are development goals, not demonstrated product capability.
- Radar-triggered PTZ cueing (R&D goal)
- Handover to control room / VMS (R&D goal)
- No cloud dependency
Ozense detects, assesses, documents and alerts. Active counter-drone measures – jamming, takeover or kinetic effects – are explicitly not part of the system.
Drone detection for sensitive and critical infrastructure
Small non-cooperative drones can pose an espionage, disruption or security risk to industrial, energy, research and other sensitive sites. Operators typically face the same three barriers: cost, lack of in-house expertise, and legal uncertainty.
Ozense is investigating a radar-first approach that detects and tracks such objects locally and, in perspective, feeds relevant events into existing security processes and control rooms. Requirements typical of critical infrastructure are taken into account during development.
Important qualification: depending on classification and risk analysis, the German KRITIS Umbrella Act may increase pressure for resilience measures and evidence. It does not create a blanket legal obligation to buy a drone radar. Ozense is neither KRITIS-certified nor approved for any particular facility type.
Radar, RF detection and cameras compared
| Capability | Radar (Ozense, R&D) | Camera / optical | RF detection |
|---|---|---|---|
| Independent of visible light | Yes | No | Yes |
| Autonomous drones with no radio link | Yes – inherent to the principle | Yes, with visibility | No |
| Range & radial velocity | Yes | Only with extra assumptions | Limited |
| Target features used for assessment | Track, motion and micro-Doppler features | Optical shape and motion features | Signal and protocol features when a radio link is present |
| Produces image data | Radar channel produces none; camera planned event-driven only | Continuous image data | No images |
| Weather and site robustness | To be validated in the field | Depends on visibility and light | Depends on the RF environment |
Radar works independently of visible light. Weather and site robustness, and performance under real clutter, are validated in the field and are not assured in general terms.
What is demonstrated today — and what is not
Demonstrated: radar-based detection and temporal tracking of moving target candidates under real environmental conditions, supported by a reproducible evaluation chain.
In validation: range, site robustness, false-alarm behaviour and robust discrimination between drones, birds and environmental clutter. Site testing and integration into existing security processes form the next pilot phase.
Assess a pilot project for your site.
Together, we define the operating scenario, relevant target types and measurable criteria for a site-specific pilot trial.
Enquire about a pilot project