Drone MANET Communications in Contested Environments
Drone operations moved from single FPV strikes to coordinated formations, and now to swarms of cooperating UAS. A swarm shares sensing and targeting data, re-plans around lost members, and saturates point-defence systems from multiple axes. All of this depends on a mesh datalink that survives jamming, node loss and a topology that changes all the time. This page describes a NetSim model of that problem. A mixed force of strike, ISR and relay drones operates over an L-band ad-hoc mesh in a 5 km × 5 km contested area. The model stress-tests the mesh with an adversary jammer and a kinetic kill on a backbone relay.
The Scenario
A drone strike and reconnaissance operation runs over a 5 km × 5 km area. The force has three groups:
- Eight FPV strike drones engage ground targets.
- Two ISR drones provide wide-area surveillance.
- Four high-altitude relay drones form a backbone that connects the forward UAS to one Ground Control Station (GCS) at the southwest edge of the area.
The adversary brings two threats: a jammer parked near the centre of the operating area, and a short-range air-defence system that can engage low-altitude UAS.
Figure 1: Initial node deployment. The shaded red circle is the 750 m effective radius of the jammer. It covers Strike-1, Strike-2, ISR-1 and Relay-2.
Force composition
8 × Strike UAS
FPV attack drones make precision strikes against ground targets. They ingress, loiter and egress in two approach groups.
2 × ISR UAS
These drones do wide-area surveillance and target acquisition. They orbit at higher altitude and also act as comms relays for the outer strike drones.
4 × Relay UAS
A high-altitude backbone connects the forward UAS to the GCS. The backbone has enough relays that it does not partition during normal flight.
1 × Ground Control Station
The GCS is stationary at the southwest edge. It is the single ingress point for command and control. It sees the whole mesh through Relay-1 and Relay-4.
1 × Jammer + 1 × Air Defence
These are the adversary assets. The jammer targets the L-band mesh. The air-defence system can engage low-altitude UAS at short range.
The Network
All fifteen nodes share one L-band DTDMA mesh with AODV routing. Connectivity is range-based at 1500 m. Any two nodes inside that radius have a link, and nodes farther apart do not. This hard cutoff forces traffic to go multi-hop through the relay backbone. It also makes mobility and node loss important, because the best next hop at one time can be gone soon after.
| Parameter | Value |
|---|---|
| Frequency band | L-band, 1350–1400 MHz |
| Channel bandwidth | 5 MHz |
| MAC protocol | DTDMA, frequency hopping enabled |
| Routing protocol | AODV (reactive, suits mobile UAS) |
| Aggregate channel capacity | 12 Mbps |
| Communication range | 1500 m (range-based pathloss) |
| TX power | 20 W per node |
| Modulation / coding | 16-QAM, rate 3/4 |
Table 1: Main radio and network parameters. The PHY values come from published specifications of fielded tactical L-band MANET radios.
Traffic Carried Over the Mesh
Four application flows model the operational traffic. Three flows go uplink to the GCS. Only the C2 stream goes the other way.
| Flow | Direction | Rate | What it is |
|---|---|---|---|
| C2 commands | GCS → each Strike UAS | 64 kbps × 8 | Joystick and waypoint updates, with a latency budget below 200 ms. |
| FPV video | each Strike UAS → GCS | 64 kbps × 8 | Low-bitrate H.265 from the strike camera. |
| PLI | each UAS (all 14) → GCS | 16 kbps × 14 | Position / status reports for situational awareness. |
| ISR imagery | each ISR UAS → GCS | 128 kbps × 2 | Wider field-of-view surveillance video. |
Table 2: Application flows. The aggregate offered load is about 1.5 Mbps, and the channel carries 12 Mbps. Hop count limits the mesh, not bandwidth.
The Stress Events
The 70-second simulation includes two adversary actions. Each stresses a different failure mode: one is electronic and one is kinetic.
Figure 2: Stress event timeline. Both events affect the same mesh.
Event 1: L-band jammer activation (t = 30–60 s)
A jammer at (2200, 2500) jams the 1350–1400 MHz band. Drones inside its 750 m radius see a 20 dB rise in the noise floor. The SINR then drops enough to lose packets, although the link still “exists” by range. At the initial node positions, the jammer affects Strike-1, Strike-2, ISR-1 and Relay-2. Relay-2 is one of the two relays that carry the central backbone.
Event 2: Relay-4 shoot-down (t = 45 s)
Fifteen seconds into the jamming window, adversary air defence engages and destroys Relay-4. The kinetic loss is instantaneous and permanent, with no warning, no graceful shutdown and no SINR ramp. Relay-4 is one of the two backbone neighbours of the GCS. After its loss, every remaining packet must enter the mesh through Relay-1, while Relay-2 is already jammed.
Three Experiments
Three experiments add the two stress events one at a time, so that each failure mode shows separately.
| Experiment | Configuration | What it isolates |
|---|---|---|
| Exp 1: Baseline | No jamming, no failure | Best-case PDR, latency and PHY-error floor under nominal mobility. |
| Exp 2: Jamming | L-band jammer active 30–60 s | Localized PHY-layer impact on the four drones inside the jammer zone. |
| Exp 3: Jam + Shoot-down | Exp 2 + Relay-4 destroyed at t = 45 s | How far damage spreads when a backbone relay is lost during EW. |
Table 3: Experiment matrix. We ran each experiment with 4 seeds and report 95% confidence intervals.
Results
Three indicators show the stress increase across the experiments:
- The PDR of the most-jammed flow: FPV from Strike-1, the strike drone closest to the centre of the jammer.
- The PHY-layer error count. It shows the jammer effect directly, even when frequency hopping protects the PDR.
- The number of nodes whose performance each stress event materially affects.
Figure 3: Three indicators of progressive stress. Left: PDR for the FPV flow from Strike-1. Middle: PHY-layer error count over the 70 s window. Right: number of materially affected UAS. In Exp 2, these are the nodes inside the jammer zone. In Exp 3, they also include the nodes whose primary route to the GCS used Relay-4.
What the results say
- Frequency hopping absorbs the jammer at the PDR layer. Strike-1 is inside the 750 m circle of the jammer and still delivers 88% of its FPV packets, less than 2 points below baseline. The jammer causes a 263× rise in PHY errors, but the DTDMA waveform protects most of the user traffic.
- Jamming is geographically contained. In Exp 2, only the four UAS inside the jammer circle show any degradation. The other ten continue at baseline.
- The kinetic loss spreads the damage. The loss of Relay-4 takes one of the two backbone neighbours of the GCS offline. Every remaining route must then go through Relay-1. The number of materially affected nodes jumps from 4 to 12, including drones nowhere near the jammer.
- The mesh stays up. Through both stress events, every flow continues to deliver some traffic. The network is degraded but functional, which is the property that a tactical mesh must have.
Why It Matters
This combination of mobility, coordinated electronic warfare and a kinetic kill on a key node is almost impossible to test on real hardware. In NetSim, it is a configuration file. The same scenario template extends to other routing protocols, different waveforms, larger swarms, mobile jammers, multi-jammer geometries and progressive air-defence engagements. Use this kind of analysis to evaluate a tactical MANET radio, design a swarm comms architecture, or study resilience in contested environments.
See this scenario running in NetSim
Walk through the configuration, the trace logs, and the result extraction with one of our engineers.