How a 5-Node Mesh Detected 23 Simultaneous Drone Incursions at a Border Sector

Context and challenge

A coastal border enforcement team responsible for a 25 km² sector faced a rapidly evolving problem: small unmanned aircraft operating close to shorelines and crossing points, often flying low, intermittently, and in groups. Conventional surveillance in the sector relied on a patchwork of tools—visual observation, fixed cameras, and radar coverage that was not optimized for small, low-signature targets moving against sea clutter and variable weather.

The team’s operational goal wasn’t simply “detect a drone.” It was to detect, track, and maintain identity across multiple targets simultaneously—without overwhelming operators with false alarms or losing tracks when drones moved behind terrain features, infrastructure, or maritime clutter.

A live exercise was planned to stress-test readiness against a swarm-style incursion. The scenario posed several specific challenges:

• High target density: multiple drones entering the sector at once, forcing the surveillance system to maintain many simultaneous tracks.
• Low-altitude flight near water: reflections, wave clutter, and atmospheric ducting can mask or distort target signatures.
• Intermittent visibility: targets may appear and disappear due to line-of-sight limitations, infrastructure, or flight behavior.
• Operator workload: even a modest number of drones can exceed manual tracking capacity if the system does not manage correlation and prioritization.
• Border constraints: deployments must be fast, resilient, and able to operate across mixed terrain (shoreline, built-up areas, and open coastal stretches).

The exercise’s success criteria focused on two outcomes: reliable detection and continuous tracking across the sector, and actionable accuracy sufficient to support interception planning and evidence-grade reporting.

Approach and solution

To address these constraints, the border team deployed a 5-node AISAR Grid arranged as a mesh across the coastal sector. Each node contributed sensing and processing into a shared picture, designed to maintain performance even when individual nodes faced temporary occlusions or local interference.

Why a 5-node mesh

A single sensor can detect drones under favorable conditions, but swarm scenarios expose the limitations of solitary coverage:

• Tracks can be lost when a target moves behind a building, a ridge, or other obstructions.
• Classification confidence can fluctuate if the target’s aspect angle changes or if the environment introduces clutter.
• A single node can become a bottleneck if too many targets demand simultaneous attention.

A multi-node mesh changes the geometry of detection. With several spatially separated nodes, the system can:

• Observe targets from multiple angles
• Improve localization through cross-node correlation
• Sustain tracks when one node’s line of sight degrades
• Reduce ambiguity by validating detections across the grid

Deployment design (sector-focused, not infrastructure-heavy)

The deployment prioritized coverage continuity over maximizing range from any one location. Nodes were positioned to create overlapping fields across key ingress routes—coastal approaches, nearshore corridors, and likely crossing points—while ensuring the mesh could withstand partial outages.

Operational requirements included:

• Rapid setup: minimal dependence on permanent towers or fixed infrastructure
• Resilience: continued operation if a node temporarily lost visibility, power stability, or local communications
• Low operator burden: automated correlation and track management, minimizing manual intervention

Data fusion and track management

In swarm conditions, the hardest problem is not detection; it is track continuity. Multiple targets can cross paths, change speed, or fly in formations that confuse naïve trackers. The grid’s workflow centered on:

1. Initial detection at the node level
   Each node produced candidate detections based on its local sensing returns and internal filtering.

2. Cross-node correlation
   Detections were compared across nodes to confirm target validity and reduce the chance of localized false positives.

3. Fusion into a unified track picture
   The system merged correlated detections into stable tracks, prioritizing continuity and identity persistence.

4. Accuracy refinement
   By combining multiple observation points, the grid tightened localization—supporting the exercise goal of meter-level accuracy for tracked targets.

5. Operator-facing prioritization
   Rather than presenting raw detections, the grid emphasized tracks, confidence, and motion patterns—what operators need for response decisions.

Exercise integration

The live exercise was designed to resemble operational reality: multiple drones entered the sector within overlapping time windows, with varied altitudes and routes. This ensured the test measured the system’s behavior under:

• concurrent tracking load,
• nearshore environmental noise,
• and track handoffs as drones moved across the mesh footprint.

Results

During the live swarm-incursion exercise, the 5-node grid detected and tracked 23 simultaneous targets across the 25 km² coastal sector, maintaining meter-level accuracy in its track outputs.

Key performance outcomes observed during the exercise included:

• Simultaneous multi-target tracking:
  The grid sustained 23 concurrent tracks without collapsing into track swapping or unmanageable operator alerts.

• Actionable localization:
  Meter-level accuracy enabled practical response planning—supporting intercept vectors, predicted pathing, and clear demarcation of the border sector boundaries relative to target movement.

• Improved continuity through overlap:
  Tracks remained stable as targets moved through different coverage zones, indicating that overlap and fusion effectively mitigated line-of-sight losses at individual nodes.

• Reduced false-alarm pressure:
  Cross-node confirmation and fusion helped keep the operator view focused on validated targets rather than noise—critical when many targets appear at once.

• Operationally relevant tempo:
  The system’s outputs remained usable in real time, helping teams make decisions based on evolving track behavior rather than after-the-fact analysis.

While exercise conditions can never capture every real-world complication, the outcome demonstrated that a properly arranged mesh can convert a swarm event from an “overwhelm the screen” situation into a manageable operational picture.

What made the difference

Several design choices contributed directly to the exercise outcome.

1) Geometry over brute force

Instead of relying on a single high-power sensor, the deployment used spatial diversity. Multiple nodes observing from different positions made it harder for targets to exploit blind spots and reduced the chance that environmental clutter would dominate the full picture.

2) Fusion built for density

Swarm events create classic tracking pitfalls: crossing paths, temporary merges, and sudden accelerations. The grid’s track management emphasized identity persistence, allowing operators to trust that Track A remained Track A—even when targets maneuvered near each other.

3) Accuracy that supports decisions

Meter-level accuracy mattered because it transformed detection into intervention readiness. When location uncertainty is too large, response assets can be misdirected, and enforcement actions can become risky or ineffective. Tight localization made the tracks operationally meaningful.

4) Operator workload as a first-order requirement

In high-tempo scenarios, the limitation is often human bandwidth. The grid’s ability to present a consolidated, validated picture reduced cognitive load—an essential feature when the number of targets rises quickly.

Key takeaways

• A mesh of five nodes can provide sector-wide resilience in coastal border environments, where line-of-sight and clutter frequently disrupt single-sensor coverage.
• Swarm readiness depends on track continuity, not just detection. Systems must maintain identity through crossings, maneuvers, and intermittent visibility.
• Meter-level accuracy changes the operational equation, enabling response planning and clearer enforcement decisions rather than merely reporting “something is there.”
• Cross-node correlation helps control false alarms—especially important near water and built infrastructure where clutter and reflections can mimic small targets.
• Exercise design should stress concurrency. Testing with many simultaneous targets reveals failure modes that never appear in single-drone trials.

In environments where drone incursions can arrive in groups and exploit coastal complexity, a multi-node grid approach demonstrated that maintaining a stable, accurate track picture at swarm scale is achievable—turning a potentially chaotic incident into a structured, actionable operating picture.