How drones navigate when satellites go dark, and why ants may hold the answer to the most critical capability gap in modern warfare
Since the opening months of Russia's full-scale invasion of Ukraine, Russian electronic warfare systems have maintained persistent GPS jamming and spoofing across vast swaths of the battlespace. Systems including the Pole-21, R-330Zh Zhitel, and Krasukha-4 create denial zones spanning 100 kilometers or more, rendering satellite navigation unreliable or completely unavailable for Ukrainian drone operations.[1]
The impact has been immediate and severe. First-person view (FPV) drones — which became the war's signature weapon — initially relied on GPS for waypoint navigation, return-to-home functions, and geolocation of targets. In contested EW environments, these drones lose position awareness, fly erratically, or are captured by spoofing signals that redirect them to Russian-controlled coordinates.[2]
Ukraine's response has been adaptive and rapid. Operators shifted to purely visual piloting for FPV strikes, while longer-range autonomous missions increasingly rely on inertial navigation systems (INS), visual odometry, and terrain-matching algorithms. The conflict has become a live laboratory for GPS-denied navigation — solutions that work in Ukrainian skies are assessed to define the next generation of autonomous systems globally.[3]
The most mature GPS alternative uses onboard cameras to track visual features and estimate position by matching what the drone sees against stored satellite imagery or terrain elevation models. Palantir's Visual Navigation (VNav) system, deployed in late 2025, enables drones to navigate using "onboard cameras and compute" alone — no GPS, no external signals.[4] The approach mirrors how humans navigate: recognizing landmarks and tracking movement relative to known features. Limitations include degraded performance in featureless terrain (desert, ocean) and at night without infrared.
INS uses accelerometers and gyroscopes to track movement from a known starting position — a technique dating to WWII-era V-2 rockets. Modern MEMS-based INS units are small enough for tactical drones, but drift accumulation remains the fundamental challenge: without periodic position fixes, errors compound over time. Advances in quantum inertial sensors (cold-atom interferometry) promise dramatically reduced drift, but remain laboratory-stage technology.[5]
Rather than relying on dedicated navigation satellites, this approach uses ambient radio signals — cellular towers, TV broadcasts, Wi-Fi access points, even enemy radar emissions — to triangulate position. The electromagnetic environment itself becomes the navigation infrastructure. This approach is inherently resilient to GPS-specific jamming because it exploits the adversary's own emissions as navigation aids. Zainar's RF positioning technology is assessed to operate on similar principles.[6]
Desert ants (Cataglyphis) navigate across featureless terrain using path integration — counting steps and tracking direction relative to polarized sunlight — combined with Earth's magnetic field as a compass reference. Migratory birds use cryptochrome proteins in their retinas to literally see magnetic field lines. Both biological systems operate without any external signal infrastructure.[7] DARPA's INSPIRE program and academic research at multiple universities are translating these biological navigation algorithms into drone-compatible systems. The key insight: nature solved GPS-denied navigation hundreds of millions of years before GPS existed.
When individual drones cannot determine their absolute position, a swarm can maintain relative positioning through inter-drone ranging (UWB radio, optical) and distribute any available position fixes across the network. If one drone in a hundred obtains a reliable fix — from a brief GPS window, a visual landmark match, or a signals-of-opportunity solution — the entire swarm can update. This distributed approach degrades gracefully rather than failing catastrophically.[8]
The convergence between biological navigation research and autonomous systems engineering is not metaphorical — it is direct and technical. The same research groups studying ant magnetoreception and path integration are publishing in both biology journals and IEEE robotics conferences.
Cataglyphis desert ants forage across hundreds of meters of featureless terrain and return to their nest in a straight line. They accomplish this through path integration: continuously tracking their heading (using polarized skylight patterns) and distance traveled (using a step counter), then computing a return vector. This is functionally identical to inertial navigation — but implemented in a brain with fewer than 250,000 neurons.[7]
Roboticists have implemented ant-inspired path integration in small drones and ground robots. The AntBot project at Aix-Marseille University demonstrated a six-legged robot that navigates using only a UV-light polarization compass and an optical flow sensor — no GPS, no map, no external signals. The system achieved navigation accuracy within 1% of distance traveled, outperforming consumer-grade GPS in some conditions.[9]
Migratory birds detect Earth's magnetic field through quantum effects in cryptochrome proteins — radical pair mechanisms that modulate visual perception based on magnetic field orientation. This gives them a compass that requires no infrastructure, cannot be jammed (at current technology levels), and functions globally.[10]
Translating magnetoreception into drone hardware faces challenges: Earth's magnetic field is weak (~50 microtesla), local anomalies are common near metal structures, and current magnetometers lack the sensitivity of biological systems. However, advances in quantum magnetometry — including nitrogen-vacancy center diamond sensors — are closing the sensitivity gap. A drone equipped with a biological-grade magnetometer would have a navigation reference that no electronic warfare system could deny.[10]
GPS denial and GPS-denied navigation are locked in an accelerating co-evolutionary arms race. Each advance in jamming capability drives innovation in navigation alternatives, which in turn drives more sophisticated electronic warfare responses.
The Wraith system — a 32-pound airborne EW platform introduced in late 2025 — represents the latest iteration: a drone specifically designed to hunt hostile emitters in GPS-denied zones, navigating without the very signals it is designed to disrupt.[11] This self-referential capability — an EW platform that operates in the environment it creates — is assessed to be the defining characteristic of next-generation autonomous weapons.
The endgame of this co-evolution is a battlefield where no electromagnetic signal can be trusted. GPS, cellular, Wi-Fi — all potential navigation aids become potential deception vectors. The systems that operate in this environment will be those that can navigate using physics that cannot be spoofed: inertial forces, gravity gradients, magnetic fields, and optical terrain matching. Biology arrived at this conclusion independently. Engineering is catching up.
GPS dependence is the single largest vulnerability in Western autonomous systems doctrine. The Ukraine conflict has demonstrated that any peer or near-peer adversary can deny GPS across operationally significant areas using commercially available electronic warfare technology. The assumption of satellite navigation availability — baked into decades of military planning, procurement, and training — is assessed to be fundamentally invalid for contested environments.
The solutions are emerging in parallel: visual odometry for medium-range operations, INS for short-duration missions, RF signals of opportunity for urban environments, and bio-inspired approaches for the most demanding scenarios. No single technology replaces GPS. The future of autonomous navigation is multi-modal — fusing multiple imperfect sources into a reliable position estimate, exactly as biological organisms have done for hundreds of millions of years.
The strategic winner of the GPS-denied navigation race will be the force that fields autonomous systems capable of operating in electromagnetically contested environments without degradation. Based on current trajectories, this capability is assessed to reach operational maturity within 3-5 years for visual-inertial systems, and 7-10 years for bio-inspired quantum navigation. The ants, as usual, are already there.
In modern warfare, GPS cannot be relied upon.