6-layer sensor fusion for full-spectrum
drone detection in any environment.
EO/IR · Acoustic · RF · Radar · 5G · Satellite.
No blind spots. No gaps.
MARS is a multi-layer drone detection system that fuses six independent sensing modalities — visual, acoustic, RF, radar, urban wireless, and satellite — into a single unified picture for military and critical-infrastructure operators.
Each layer captures what the others miss. A drone running silent in RF (autonomous flight) is still visible to radar and EO/IR. A drone flying low in urban clutter is picked up by 5G anomaly detection and acoustic triangulation.
The MARS fusion engine correlates all six data streams in real time, producing a single high-confidence track with classification, velocity, and intent scoring — delivered to any C2 interface.
Explore the Layers →Day/night electro-optical and infrared cameras provide passive, silent detection with zero RF emissions. AI-driven object detection classifies drone type, size, and flight mode in real time. Thermal signatures expose warm motors even in full darkness or adverse weather. Gimbal-mounted optics deliver continuous tracking at ranges up to 3 km with automatic slew-to-cue from other layers.
An array of distributed microphones captures the characteristic rotor blade signature of multi-rotor drones. Time-difference-of- arrival (TDOA) algorithms triangulate the 3D position of any acoustic source with sub-10-metre accuracy. Machine-learning classifiers distinguish drone species from birds, aircraft, and background noise — even in urban soundscapes. Effective at ranges up to 500 m with no RF dependency.
Wideband software-defined radio (SDR) receivers continuously scan the 2.4 GHz, 5.8 GHz, and 900 MHz ISM bands used by consumer and commercial drone control links. Protocol fingerprinting identifies the specific remote-controller brand, drone model, and pilot location when a data-link is active. Works against manually piloted threats; complements radar and acoustic layers for autonomous or silent drones.
Dedicated UAS-mode radar detects small, low-RCS targets moving at drone speeds (0–200 km/h) and discriminates them from birds and weather using micro-Doppler spectral analysis. The rotating blades produce a characteristic Doppler spread that uniquely identifies multi-rotor platforms at distances up to 5 km. Works in all weather, day or night, regardless of RF silence or visual conditions.
In dense urban environments where radar has limited line-of-sight, MARS leverages existing 5G and WiFi infrastructure as a passive bistatic network. Movement anomalies between base stations create unique multipath signatures that pinpoint airborne objects between buildings. Additionally, drones using civilian WiFi telemetry or 4G LTE control links are passively detected via protocol inspection without emitting any signal.
In research, MARS is exploring the use of Starlink and LEO satellite constellations as passive illuminators for bistatic radar. The downlink beams of these satellites cover near- continuous global surface areas; a ground-based receiver network exploits signal reflections off airborne objects. This layer would provide wide-area detection far beyond line-of-sight — an open research frontier.
An extended Kalman filter correlates tracks from all six layers into a single probabilistic state vector. Each layer contributes confidence-weighted position, velocity, and class estimates — resolving conflicts and filling gaps in real time.
On-edge neural networks classify drone species from EO/IR imagery, acoustic spectrograms, and RF protocol signatures. Models are continuously updated via federated learning from deployed systems — no cloud dependency in theatre.
A single-screen C2 interface displays all fused tracks, classification confidence, layer contributions, and cueing recommendations. Compatible with existing military C2 networks via LINK-16, ASTERIX, and REST API.
The MARS platform is sensor-agnostic. New layers, sensors, or AI models are hot-plugged without system downtime. Deployable on bare-metal, containerised, or air-gapped infrastructure depending on operational requirements.
Perimeter protection for FOBs, command posts, and logistics nodes. Rapid-deploy sensor kits with sub-60-minute setup time.
Permanent 24/7 surveillance of power plants, refineries, and water facilities. Integrated alarm and effector cueing without operator fatigue.
Dense urban deployments leveraging existing 5G infrastructure. Low-visibility mobile units for VIP protection and public gathering security.
Ship-borne and shore-based configurations protecting naval assets and port infrastructure from surveillance and payload-carrying UAS threats.
MARS Systems brings together engineers, physicists, and veterans with backgrounds across radar signal processing, RF systems, computer vision, embedded AI, and operational military experience. Our team has designed sensors that fly, float, and fire — now unified in a detection platform where no drone goes unnoticed.