Aireon is operator of the space-based Automatic Dependent Surveillance-Broadcast (ADS-B) air traffic surveillance system. The acquisition of Aireon is a defining step in Iridium’s strategy to provide the foundational architecture for global aviation safety, bringing space-based surveillance, safety communications, PNT and operational data together on a single network.

“Aireon has always been part of Iridium’s aviation safety strategy. We founded it in partnership with the world’s leading Air Navigation Service Providers (ANSPs), because we believed space-based aviation safety was a generational opportunity,” said Matt Desch, CEO, Iridium. “The aviation industry is now entering an era of growing air traffic, denser airspace, autonomous aircraft, and greater expectations for safety and resiliency. Bringing Aireon fully inside Iridium better positions us to build what’s needed to support the future of aviation, including more innovations like the future introduction of space-based VHF communications.”

Platform for Aviation Safety
The acquisition unites Aireon’s surveillance and data services, including GPS jamming and spoofing detection, with Iridium’s global satcom network and PNT services that help keep GPS-dependent systems working in contested environments. This combination creates one company providing four critical aviation industry capabilities: knowing where every aircraft is, communicating with the pilots flying them, providing the navigation and timing integrity those aircraft rely on, and translating that information into operational insights that make airspace safer and more efficient. No other satellite operator delivers this combination of capabilities on a global scale.

The Aireon system, which is certified by theEuropean Union Aviation Safety Agency (EASA), flies as a payload on the Iridium satellite constellation and tracks an average of 190,000 flights per day. Commercial aircraft broadcast information such as an aircraft’s identity, location, altitude, speed, and heading. Aireon’s space-based ADS-B payload captures this information in real time, with 100% global coverage. ANSPs covering more than 50% of the global airspace rely on Aireon data to create safer and more efficient airspace.

The world’s leading ANSPs and investors in Aireon, including NAV CANADA and NATS (United Kingdom), AirNav Ireland, ENAV(Italy), and Naviair (Denmark), each played a vital role in launching the Aireon service, proving its reliability, and establishing it as a critical part of the global air traffic control infrastructure. NAV CANADA and NATS, which together manage the most heavily trafficked oceanic airspace in the world — the North Atlantic Tracks between Europe and North America, were the first to go live with the service. In connection with the acquisition, both ANSPs will sign extended data services agreements through 2035 and beyond, with provisions for continued cooperative development of space-based VHF communications and other new capabilities.

“Aireon and Iridium have been partners since day one, and that partnership is the reason we have been able to build the world’s only space-based air traffic surveillance system and a fast-growing aviation data services business alongside it,” said Don Thoma, CEO of Aireon. “Becoming part of Iridium is a natural next step for our team, our customers, and our roadmap, particularly as our data products expand into new areas like turbulence detection and aviation data analytics. Together, we are building the foundation for the future of global aviation.”

“NAV CANADA is proud of our foundational role in establishing Aireon’s world-first technology,” said Mark Cooper, President and CEO, NAV CANADA. “This sale sharpens our focus on our core expertise: keeping Canada’s skies safe. As a fellow founding partner, Iridium is the ideal owner to guide Aireon’s continued commercial growth. We wish the entire team continued success and look forward to our ongoing relationship as a customer.”

“We have been proud to be a part of Aireon’s successes, most notably making real-time aircraft surveillance over the Atlantic a reality for the first time in history, enabling even safer operations across the North Atlantic,” said Martin Rolfe, CEO, NATS. “As a shareholder for the past eight years, it is now the right time for us to divest. We are confident Aireon is well positioned for the future and wish the team every success in the next stage of its development.”

The Next Transition: Space-Based VHF
Space-based VHF communications represent a major opportunity in air traffic management, extending pilot-to-controller VHF services into oceanic and remote airspace where ground infrastructure cannot reach, without the need for additional aircraft equipment. The model is similar to how aircraft already carry ADS-B transceivers, which enables Aireon to deliver space-based ADS-B surveillance without requiring fleet retrofits.

Aireon’s Growing Data Services Business
Beyond surveillance for ANSPs, Aireon operates a fast-expanding aviation data services business that sells real-time and historical aviation data to airlines, airports, OEMs, governments, and aerospace operators. Product lines already available or launching this year include turbulence detection, GPS jamming and spoofing detection, and safety and efficiency analytics. Additional applications are also in development to support the rapidly evolving airspace environment.

Aireon’s data business is one of its highest-growth areas today and is expected to be a meaningful contributor to the combined company’s aviation growth.

Terms of the Transaction
Iridium is an existing owner of Aireon and will acquire the remaining 61% of equity interests of Aireon in the transaction for a purchase price of approximately $366.7 million from the other owners, NAV CANADA, AirNav Ireland, ENAV, NATS and Naviair. The purchase price will be paid 50% at closing and 50% on the one-year anniversary. Iridium will also assume Aireon’s outstanding debt, expected to be approximately $155 million at closing.

The acquisition of Aireon is accretive to Iridium’s growth outlook; over the past three years, Aireon’s total revenue has grown at a compound annual growth rate (CAGR) of 10%. Iridium expects the acquisition will result in at least an additional consolidated $100 million of service revenue and $30 million of OEBITDA on an annualized basis.

Iridium expects to pay the purchase price with current liquidity, including borrowings under its revolving credit facility, and future cash from operations. After closing the transaction, Iridium expects net leverage to increase to approximately 4.0 times OEBITDA during Q3 2026, with net leverage planned to return to the current levels over the subsequent twelve months. Iridium’s long-term net leverage guide of 2.0 times OEBITDA by the end of the decade remains unchanged and assumes no change in its paused share buyback program.

Aireon will continue business-as-usual operations in the near term, with no planned changes to business strategy. The transaction is targeted to close in early July.

Evercore served as financial advisor and Cooley and Milbank served as legal counsel to Iridium. PJT Partners served as financial advisor and Hogan Lovells served as legal counsel to Aireon.

<p>The post Iridium to acquire Aireon to lead aviation safety first appeared on GPS World.</p>

The BeiDou industry includes remote sensing and geographic information systems (GIS), mobile communications and indoor positioning. The satellite navigation sector generated 629 billion yuan (US$92 billion) in 2025, up 9.24% year on year, according to the report.

China has established a complete BeiDou industrial chain and supply chain, covering chips, modules, antennas, terminals, system integration and application services, , according to the report. Domestic capabilities are becoming increasingly self-reliant, with the cumulative shipments of BeiDou-compatible chips and modules reaching hundreds of millions, supporting a secure and robust industry supply chain.

Domestic sales of BeiDou-enabled terminals exceeded 410 million units in 2025, with more than 2.2 billion BeiDou-capable devices in use across the country.

Internationally, BeiDou services and related products have been exported to more than 140 countries and regions.

<p>The post China’s BeiDou-charged navigation industry reached $195B in 2025 first appeared on GPS World.</p>

Onocoy, a decentralized GNSS reference station network, is launching Loop Back, a new platform feature that routes quality-assured RTK correction data back to each station operator’s own devices free of charge. More than 7,800 active reference stations contribute to the onocoy network.

Operators who also needed precision positioning for their own drones, survey rovers, precision agriculture equipment, or autonomous machinery face a common friction point: the reference station they owned and operated produces valuable correction data, but routing that data back to their own field equipment requires either a separately maintained NTRIP caster or an additional subscription. Loop Back eliminates both.

Loop Back is immediately available to all onocoy station operators as a standard platform feature. Full documentation and setup guides are available at docs.onocoy.com.

How Loop Back works

When a GNSS reference station is connected to onocoy, raw observation data flows from the operator’s hardware into onocoy’s quality validation pipeline. The platform continuously checks position stability, multi-constellation health (GPS, GLONASS, Galileo, BeiDou), uptime and other parameters before producing a quality-assured RTCM 3 correction stream.

That validated stream has two destinations simultaneously: enterprise data clients who purchase GNSS reference station data through onocoy’s pay-per-use model, and the station operator’s own devices via Loop Back. The operator receives the same production-grade correction stream used by commercial clients, free of charge and with no data credits consumed.

Key capabilities at launch:

• Up to three simultaneous active connections from an operator’s own devices to their own station’s corrections, with unlimited devices configurable
• Compatible with any NTRIP-capable station regardless of hardware brand or model
• Quality monitoring identical to that applied to enterprise client streams
• No separate NTRIP caster required; onocoy manages the infrastructure

• Free of charge: No data credits consumed for the operator’s own station data.

Who benefits

Loop Back is designed for the growing segment of professionals who both operate a reference station and rely on precision positioning in their daily work. Target use cases include:

• Precision agriculture: Farmers running auto-steered machinery, UAV-based crop monitoring, and variable-rate application systems
• Geomatics and surveying: Professionals running a base station and multiple rover units across a site, eliminating the overhead of a local base-rover setup
• Autonomous systems, robotics and drones: Operators deploying multiple vehicles or aircraft requiring cm-accurate positioning for mapping, inspection, or delivery workflows
• Research: Academic and scientific teams running parallel measurement campaigns from a shared base station.

Economics of station operation

Most professionals who deploy a GNSS reference station do so because their business in precision agriculture, surveying, drone operations and construction demands one. By connecting that station to onocoy, operators put the same hardware to work a second time: contributing data to onocoy’s global network and earning rewards worth several hundreds of U.S. dollars per year.

That additional income is enough to amortize the station in under two years before accounting for potential savings on subscriptions. Because onocoy applies continuous quality monitoring to every stream, operators also safeguard the positioning accuracy their business depends on.

<p>The post Onocoy’s Loop Back gives reference station operators RTK corrections for their devices first appeared on GPS World.</p>

This article presents the first direct comparison of Cyclone Global Navigation Satellite System (CYGNSS) — a NASA GNSS reflectometry constellation — and NASA-ISRO Synthetic Aperture Radar (NISAR) — an L-band synthetic aperture radar for GPS jammer localization. The results challenge assumptions about which modality performs better and reveal that the answer depends on a question most analysts forget to ask.

The setup: Known jammer, known position

Validation requires ground truth. With help from the PNT community, we identified a GPS jammer operating near 27.32°N, 52.87°E (approximately 50 km southwest of Shiraz) that was active on Jan. 8 and Jan. 20, 2026, with confirmed quiet periods on Dec. 15 and Dec. 27, 2025. The jammer’s position was established through independent signals intelligence.

This gave us a controlled experiment: two “jammer ON” dates and two “jammer OFF” baseline dates, with satellite coverage from both CYGNSS and NISAR spanning the full period.

Two satellites, two physics

CYGNSS is a constellation of eight microsatellites that measure GPS signals reflected off Earth’s surface. Each spacecraft carries a delay-Doppler receiver that maps reflected signal power across a grid of delay and Doppler bins, known as the delay-Doppler map, or DDM. When a terrestrial jammer is active, it floods the GPS band with noise, elevating the DDM noise floor and suppressing the coherent surface reflection. The effect is detectable hundreds of kilometers from the jammer, creating a wide-area footprint in the reflected signal data.

NISAR operates an L-band SAR at 1.257 GHz, just 30 MHz from the GPS L2 frequency at 1.2276 GHz. When a GPS jammer’s broadband emissions leak into NISAR’s receive band, they create characteristic streaks in the SAR imagery. The streaks are elongated in the cross-track (range) direction, not along-track, a counterintuitive result that follows directly from SAR signal processing. In azimuth (along-track), the jammer is a fixed-point source with a valid Doppler history, so the SAR azimuth processor focuses it correctly, similar to any ground target. But in range (cross-track), the jammer’s broadband noise does not match the SAR’s chirp waveform, so range compression smears the energy across many range bins rather than compressing to a point. The result is a streak perpendicular to the flight direction, whose along-track centroid encodes the jammer’s latitude and whose cross-track extent encodes a range arc, which is the distance from the orbit ground track (FIGURE 1). The bearing of each streak encodes the jammer’s direction relative to the satellite’s ground track.

The two sensors could hardly be more different. CYGNSS sees the jammer’s effect on reflected GPS signals, offering an indirect measurement spread across hundreds of specular reflection points. NISAR sees the jammer’s emissions directly in its own receiver, which is a more precise measurement, but only along the satellite’s narrow ground track. FIGURE 2 shows both detection sets converging on the jammer location.

CYGNSS: 785 Detections, 4.33 km Error

We processed all CYGNSS Level 1 data within 200 km of the jammer location on both ON and OFF dates. Four detection methods contributed observations:

■ DDM noise floor (419 detections): The pre-computed ddm_noise_floor variable, calibrated against the thermal noise reference, proved the strongest discriminator. Near-jammer values exceeded 15,000 counts against a ~10,000 mean background.

■ Spatial noise grid (299):A 10 km gridded analysis identified cells with anomalously elevated noise relative to adjacent cells.

■ SNR hole detection (66): Coherent surface reflections were suppressed near the jammer, creating spatial “holes” in the SNR field.

■ NBRCS drop (1): Surface reflectivity dropped approximately 16% near the jammer, though this method produced few threshold exceedances.

Across four DDM channels per spacecraft and multiple passes, this yielded 785 total anomalous observations on the jammer-ON dates.

Localizing using a simple centroid of all 785 detection positions placed the jammer 32.1 km from truth, with too many distant, low-SNR detections diluting the estimate.

Instead, we fit a parametric 1/r² inverse-distance model:

I(r)=Ar2

where A is a free amplitude parameter and r is the distance from a candidate jammer position. We jointly optimized the jammer position and amplitude using SciPy’s Nelder-Mead optimizer across all 785 observations, weighted by intensity. The optimizer converged on a position 4.33 km from ground truth, providing a 27.7 km improvement over the centroid (FIGURE 3).

The baseline: Zero false positives

On the jammer-OFF dates (Dec. 15 and Dec. 27, 2025), the pipeline produced exactly zero detections using the same thresholds, geographic area and satellites: a completely clean result. This suggests that the 785 detections are unlikely to be sensor artifacts or geographic anomalies. They disappear when the jammer turns off.

NISAR: 17 Detections, 6.26 km Error

NISAR’s approach is fundamentally different. Rather than measuring hundreds of reflected signals across a wide area, it captures direct emissions in a narrow swath, but with far greater geometric precision.

We processed NISAR L2 GCOV (geocoded covariance) products from Track 157, Frame 15 (ascending) for three dates: the Dec. 27 baseline and the Jan. 8 and Jan. 20 jammer-ON passes. The detection pipeline used eigenvalue decomposition of the polarimetric covariance matrix:

1. λ₁ ratio thresholding: In jammer-contaminated pixels, the dominant eigenvalue λ₁ of the 2×2 [HH, HV] covariance matrix rises sharply relative to the scene mean, indicating an unpolarized additive source.
2. Cross-polarization ratio (HV/HH): GPS jammer emissions are unpolarized, disproportionately elevating the HV channel. Anomalous HV/HH ratios flag contaminated azimuth lines.
3. Iterative outlier trimming: Three rounds of 1.5σ clipping removed scattered false detections, leaving 17 high-confidence streak centroids.

With detections from two passes on different dates, we had two independent bearing lines. Each pass’s streak centroids defined an azimuth aligned cluster whose major axis pointed toward the jammer. A PCA fit to the two clusters extracted the bearing: 308.1° from the Jan. 8 pass and 316.2° from Jan. 20. Their intersection — computed via scipy optimization of the angular residual — landed 6.26 km from ground truth (FIGURE 4).

The along-track/cross-track decomposition reveals why the 6.26 km error is a geometric ceiling for this dataset, not a processing limitation. Both passes come from the same Track 157 ascending orbit on a 12-day repeat cycle. The intensity-weighted along-track centroids land at +3.0 km and +3.1 km north of the jammer, a direct stable latitude measurement. The cross-track centroids land at +5.4 km and +5.6 km east of the orbit ground track, a range measurement. But because both passes share identical orbit geometry, the two range arcs are nearly parallel. The bearing difference between passes (308.1° vs 316.2°) is only 8.1°, producing a shallow intersection angle and poor cross-range resolution. A single descending pass, which would cross the ascending track at approximately 60-70°, would transform the geometry from two near-parallel lines to a genuine triangulation, potentially reducing the localization error to sub-2 km. Unfortunately, no descending NISAR pass covering this jammer site was available in the beta archive, which ends on Jan. 20, 2026.

The CEP (circular error probable, the radius containing 50% of repeated estimates) was 6.88 km, meaning if we ran this analysis on many similar jammers, half our estimates would fall within ~7 km.

Who wins?

CYGNSS wins, and not just on accuracy.

A naive confidence metric for the 1/r² fit would be the scatter of the 785 input detections (CEP = 127 km). But the detections are not the estimate; they are the inputs to a model fit. The relevant confidence question is: How stable is the fitted position?

We answered this with a 500-iteration bootstrap: resample the 785 detections with replacement, re-run the 1/r² optimizer each time and measure the spread of the resulting position estimates. The bootstrap CEP, the median radial distance across 500 fitted positions, was 3.48 km. The optimizer converges stably to within a few kilometers of the same location regardless of which detections are included.

This means CYGNSS achieves 4.33 km error with 3.48 km confidence, both better than NISAR’s 6.26 km error and 6.88 km confidence.

The bootstrap CEP also reveals what the raw scatter obscures: the 1/r² fit is constrained primarily by the ~80 high-intensity detections within 30 km of the jammer. The remaining 700 distant, low-intensity detections contribute little to the position estimate — they are correctly downweighted by the intensity-weighted least squares. The fit’s stability comes from the physics: a 1/r² signal has steep gradients near the source, providing strong positional constraints where it matters most.

Bayesian fusion: Can we get both?

The obvious next question: Can we combine CYGNSS’s wide-area sensitivity with NISAR’s geometric precision? We implemented four fusion strategies, all designed to work without ground truth:

■ Bayesian Gaussian posterior: Model each sensor’s estimate as a 2D isotropic Gaussian with σ = CEP/1.1774. The posterior is the product of the two Gaussians: an analytical precision-weighted mean.

■ NISAR-prior constrained 1/r²: Re-run the CYGNSS optimizer with a Gaussian regularization term pulling toward the NISAR estimate, sweeping the regularization weight λ from 0.01 to 10.

■ NISAR-proximity re-weighted 1/r²: Apply a Gaussian kernel centered on the NISAR estimate to the CYGNSS detections before fitting, effectively upweighting observations consistent with the SAR result.

■ Joint CEP-balanced: Combine the CYGNSS gradient signal with NISAR cluster proximity, weighted by (σ_CYGNSS/σ_NISAR)².

With the bootstrap CEP, the precision ratio flips. The CYGNSS Gaussian (σ = 2.95 km) is now 2× tighter than NISAR (σ = 5.84 km). The Bayesian posterior, the precision-weighted mean, lands at 4.69 km, pulling toward CYGNSS’s better estimate while incorporating NISAR’s independent geometric constraint. FIGURE 5 shows the fusion: two comparable Gaussians whose product is tighter than either alone.

The fused result (4.69 km error, 7.85 km CEP) is not quite as accurate as CYGNSS alone (4.33 km), because NISAR’s 6.26 km estimate pulls it slightly away from truth. But operationally, the fusion provides a cross-validated answer: two independent physics arriving at similar locations builds confidence that neither sensor is producing an artifact.

The key insight is that the bootstrap CEP unlocked meaningful fusion. When the raw scatter CEP (127 km) was used, NISAR dominated the posterior 343:1 and fusion added nothing. With the fit-based CEP (3.48 km), both sensors contribute, and the posterior reflects genuine multi-modal evidence.

Operational implications

For CYGNSS: CYGNSS excels at both detection and localization. Its 785 detections across a 200 km radius, with zero false positives on baseline dates, provide unambiguous jammer detection. The 1/r² fit achieves 4.33 km accuracy with a bootstrap-verified 3.48 km CEP, meaning an analyst can trust the result to single-digit kilometer precision without ground truth. CYGNSS’s eight-satellite constellation also provides sub-daily revisit, enabling near-real-time monitoring.

For NISAR: NISAR provides independent geometric confirmation. With just two passes over an active jammer, the bearing intersection achieved 6.26 km accuracy with a 6.88 km CEP. The 6.26 km result is constrained by orbit geometry, not by detection sensitivity. Our two ascending passes from Track 157 produced nearly parallel range arcs with only 8.1° of bearing separation. Adding a single descending pass would provide a crossing angle of 60° to 70° and could reduce localization error to sub-2 km — transforming NISAR from a confirming sensor into a precision localization tool in its own right. The limitation in this study was data availability: The NISAR beta archive contained only ascending Track 157 passes over the jammer site. NISAR’s 12-day repeat cycle and fixed ground track also mean the jammer must be active when the satellite passes overhead. NISAR’s current value is as a confirming sensor — when both modalities converge on the same location, confidence increases beyond what either achieves alone.

For Fusion: With comparable CEPs (3.48 km vs 6.88 km), fusion now produces genuinely blended estimates. The Bayesian posterior at 4.69 km reflects real multi-sensor information. Future improvements, such as more NISAR passes with diverse bearings or CYGNSS multi-week accumulation, would tighten both estimates further.

For the Adversary: These results demonstrate that GPS jammers operating in contested airspace are observable and localizable from orbit using openly available civilian satellite data. The 4.33 km CYGNSS result is approximately 2× better than the published state of the art for GNSS-R jammer localization (~9 km grid resolution, Chew et al., 2023) and the NISAR bearing intersection approach has not been previously demonstrated for jammer geolocation.

Still broadcasting: Jammer persistence through conflict

The validation analysis used January 2026 data. But on Feb. 28, armed conflict erupted in the region. Did the jammer survive?

We ran the CYGNSS noise floor detection pipeline for each day from Feb. 28 through April 6, comparing against the December 2025 baseline. The answer is unambiguous: The jammer is not only still active — it is operating at dramatically higher power.

In January, the jammer elevated the CYGNSS noise floor by approximately 15% above baseline. By early March, days after the conflict began, noise elevation had jumped to 50% to 60%. By mid-March, it reached 70% to 84%, where it remained through early April. Detection counts tell the same story: 89 to 192 per day in January, rising to 1,000 to 2,000 per day during the conflict (FIGURE 6).

The escalation was immediate. On Feb. 28, noise elevation was +34.5%, already double the January level. By March 3, it had reached +62.7%, and by April 6, it peaked at +79.1%. The signal has remained at 5× the January intensity through the most recent available data (April 6, 2026).

Several interpretations are consistent with this pattern:

■ Power increase: The operator increased jammer output power, perhaps in response to the conflict or as a defensive posture against GPS-guided munitions.

■ Additional jammers: Multiple units may have been co-located or deployed nearby, creating an aggregate signature larger than any single device.

■ Duty cycle change: The jammer may have shifted from intermittent to continuous operation.

What is clear is that the jammer we localized in January was not incapacitated by the conflict. It was amplified. CYGNSS’s sub-daily revisit capability makes this kind of persistent monitoring possible using entirely passive, civilian satellite data — no tasking, no cooperation with the target state and no risk to reconnaissance assets.

Context and prior work

CYGNSS-based RFI detection builds on work by Chew et al., 2023, who demonstrated grid-level jammer detection at approximately 9 km resolution using DDM noise floor anomalies. Our 1/r² parametric fit extends this from detection to localization, achieving sub-5 km accuracy by exploiting the physics of signal power decay.

At the other end of the precision spectrum, Murrian et al., 2021, demonstrated ~220 m jammer localization using ISS-mounted Doppler measurements of raw intermediate-frequency (IF) data. This approach achieves an order of magnitude better precision than our methods but requires specialized hardware and raw signal access not available on current operational satellites.

The NISAR bearing intersection approach demonstrated here is, to our knowledge, the first published use of L-band SAR RFI streaks for jammer triangulation. The key insight is that NISAR’s proximity to GPS L2 (just 30 MHz separation) makes it an unintentional but effective GPS interference sensor.

Summary

Two satellites, two physics, one jammer. CYGNSS sees the interference footprint across hundreds of kilometers and localizes the source through inverse-distance physics. NISAR sees the emissions directly in its SAR receiver and triangulates through bearing intersection. Both achieve sub-7 km accuracy independently; together, they cross-validate and build the confidence that operational use demands.

The jammer near Shiraz is still there — louder than ever. The satellites are still watching.

<p>The post Converging on the jammer: Dual-satellite GPS interference localization from space first appeared on GPS World.</p>

The delivery integrates part of Vantor’s Raptor, a vision-based software suite that enables autonomous systems to navigate, orient and extract accurate ground coordinates without relying on GNSS, with the GXP software ecosystem, ensuring intelligence continuity when sensors are degraded.

In modern conflict zones, the proliferation of inexpensive unmanned aerial systems (UAS) with equally low-quality sensors, in addition to widespread GPS spoofing and jamming, have rendered traditional drone video collection unreliable. Significant metadata drift in tactical video feeds leads to “targeting paralysis”: high-quality imagery is available, but the underlying geographic coordinates are too inaccurate for precision activities.

To solve this, Raptor Sync georegisters the full-motion video feed from the drone’s on-board camera with Vantor’s 3D terrain data in real time, enabling downstream GXP intelligence fusion, multi-domain interoperability across different sensors, and accurate ground coordinate extraction at a demonstrated absolute accuracy of <3 m. The system enables previously impossible intelligence and targeting workflows.

“In contested environments, the sensor’s imagery and video collections are only half the battle; the accuracy of the data it produces is what determines mission success,” said Kurt de Venecia, senior director of Product Development at BAE Systems GXP. “By including Raptor directly into our GXP intelligence workflows, we are providing analysts with the ability to maintain absolute targeting confidence, even when the platform’s systems or inertial sensors lack high absolute accuracy.”

Injecting corrected key-length-value (KLV) metadata from Raptor directly into the drone’s video stream at the edge enhances accuracy prior to exploitation in GXP software. This overrides inaccurate telemetry, enabling analysts using GXP solutions to extract weapon-quality coordinates and execute intelligence and targeting missions in real time.

“Analysts cannot afford to lose confidence in where a target actually is,” said Paul Millhouse, senior director ofRaptor Products at Vantor. “By using Raptor to correct video before it enters the GXP Ecosystem, we’re enhancing the performance of existing and new drone fleets. The result is a more resilient workflow for extracting accurate ground coordinates and maintaining operational tempo.”

These capabilities will be highlighted at GXP360° Professional Exchange & Workshop in San Diego, California (May 18-20).

<p>The post BAE Systems GXP, Vantor fight EW with high-accuracy targeting for drones first appeared on GPS World.</p>

This service is available with Telit Cinterion’s dual-frequency GNSS modules and NExT cellular connectivity. IoT customers gain one source for the hardware, connectivity and Skylark Dx correction data needed for sub-meter positioning.

What began in 2024 as a technical partnership has grown into a comprehensive joint offering, uniting hardware, connectivity, and corrections into a seamless solution for IoT customers.

Telit Cinterion customers can now buy modules, connectivity and corrections under one contract. For many IoT projects, this cuts vendor coordination and avoids the cost and operational complexity of building or subscribing to an RTK base-station network.

Skylark is available in three variants — Skylark Dx, Cx, and Nx RTK — to meet a broad range of requirements for accuracy, coverage, bandwidth, and power consumption.

All Telit Cinterion dual-frequency L1 + L5 GNSS modules offer native support for Skylark Dx, which streams differential GNSS (DGNSS) corrections directly to the receiver over the cellular network. Skylark Dx runs over standard RTCM via Internet Protocol (NTRIP), using minimal bandwidth and power, and provides country-wide coverage. This makes it practical for IoT devices with limited bandwidth or tight power budgets.

Typical applications include:

• e-mobility
• fleet and asset tracking
• robotics
• drones

These are cases that don’t require centimeter-level RTK accuracy but do need reliable sub-meter positioning. Customers requiring higher accuracy can upgrade to Skylark Nx RTK on compatible module variants without redesigning their devices or changing suppliers.

“Customers tell us they want precise positioning without complexity,” said Neset Yalcinkaya, president of IoT hardware at Telit Cinterion. “We’re bundling Skylark Dx with the GNSS modules and cellular connectivity we already ship. This gives customers one supplier and a single integration approach, plus a clear path to RTK down the road.”

“At Swift Navigation, our mission is to make precise positioning a standard capability,” said Holger Ippach, chief operating officer at Swift Navigation. “This partnership advances that vision by embedding Skylark into Telit Cinterion’s GNSS modules and connectivity, giving customers direct access to reliable, sub-meter positioning without the integration overhead traditionally required.”    

Service Availability

Skylark Dx is available now with Telit Cinterion solutions in Europe, North America, Japan, South Korea and Taiwan. Coverage will expand as Swift Navigation adds regions.

For more information, visit these links:

• High-Precision GNSS: RTK and DGNSS Solutions for IoT
• How GNSS Modules Achieve Higher Precision and Why It Matters

<p>The post Telit Cinterion bundles Swift’s Skylark into integrated IoT positioning first appeared on GPS World.</p>

Trimble is supporting Project Pressure by providing advanced GNSS positioning technology and research funding for the nonprofit organization’s latest expedition to map the disappearing tropical glaciers of Puncak Jaya in Papua, Indonesia.

Project Pressure has released a centimeter-accurate, 3D model of the receding ice, created using Trimble positioning technology and drone-based photogrammetry. The model establishes a scientific baseline for calculating the rate of glacier recession and projecting the timeline of disappearance.

Puncak Jaya, the highest peak in Oceania and one of the Seven Summits, is expected to be the first of the seven continental peaks to lose its glaciers as global temperatures rise.

Local communities use the data to make informed choices about crop selection and prepare for expected water shortages caused by the loss of vital reservoirs.

This expedition marks the third successful outing in Project Pressure’s “Melting Topics” series, which focuses on mapping equatorial glaciers. Trimble provides its GNSS mapping technology and research funding from the Trimble Foundation Fund to support Project Pressure in gathering critical data in some of the world’s most remote and hostile environments.

“Mapping these glaciers before they disappear is of critical importance to establish a baseline to track the glacial regression and for the local communities to understand what is happening with their water source, allowing them to adapt to a changing climate,” said Eliot Jones, senior manager, strategy and partner development at Trimble. “Through a combination of precision technology, detailed project planning and rigorous science, the models created by Project Pressure are shared for scientific study and provide a visual reference for future generations.”

Precision under pressure in hostile terrain

Mapping glaciers at altitudes exceeding 4,800 meters (15,000 feet) presents extreme logistical and environmental challenges. Near-constant cloud cover and heavy rainfall in Papua often render satellite imagery unusable, making ground-based georeferencing essential.

The expedition team installed precise geolocation reference points directly on the glacial surface at multiple locations. Using the Trimble Catalyst DA2 GNSS system and Trimble TDC600 handheld, researchers captured the exact coordinates of those points with centimeter-level accuracy. Drone imagery was then processed against the Trimble coordinates to produce a scientifically reliable 3D model of the glacier.

“Trimble makes incredibly complex technology feel simple in the field,” said Klaus Thymann, scientist and lead explorer. “When you’re standing on a glacier in freezing conditions, wearing thick gloves and surrounded by clouds, you don’t have time to fight with equipment. With Trimble, I can capture centimeter-accurate readings and the interface is so intuitive that even someone with no prior training can help collect data. That kind of reliability and simplicity is critical when you’re working in some of the most remote and challenging environments in the world.”

This approach builds on methods developed during Project Pressure’s 2024 expedition to the Rwenzori Mountains in Uganda, which also used Trimble technology.

The lightweight Trimble Catalyst DA2 GNSS system was critical for the expedition, which required helicopter access to Basecamp, followed by a trek to the launch point.

<p>The post Trimble tech enables cm-accurate 3D model of disappearing glaciers first appeared on GPS World.</p>

Both modules feature pin-to-pin compatibility, enabling platform flexibility and simplifying product development across vehicle generations as well as jamming and spoofing detection to mitigate the impact of security risks.

The ZED-X20K is designed for mass-market ADAS L3 and TCU/IVI applications, delivering lane-level accuracy worldwide using all-band GNSS and native Galileo High Accuracy Service (HAS). By eliminating the need for paid correction services, backend infrastructure, or service management, it reduces total cost and accelerates time-to-market while maintaining consistent global performance. 

For applications that require a functional-safety concept for GNSS sensors, the ZED-A20K introduces a new architectural approach. It provides ISO 26262 ASIL-B(D)-compliant GNSS RAW data simultaneously to high-performance QM positioning outputs within a single module. This enables OEMs to transition from traditional dual hardware based-GNSS systems to a single module approach, reducing system complexity and cost. 

With flexible support of externally hosted positioning engines, especially for ADAS of Levels 3 and up, the A20 concept enables enhanced flexibility for SDV–based architectures. The form-factor compatibility between ZED-X20K and ZED-A20K allows the flexibility to equip different trim levels with or without functional safety requirements out of a single socket.

The ZED-X20K has reached the engineering sample stage, and its evaluation kit is available. Samples for the ZED-A20K will be available starting in August.

<p>The post U-blox expands auto GNSS portfolio, enabling ADAS & increasing safety first appeared on GPS World.</p>

UAV Navigation — a division of Grupo Oesía specializing in advanced guidance, navigation and control solutions for unmanned vehicles — has launched the Vector-300high‑performance autopilot.

Vector-300 is designed to meet the industrial and operational requirements of mass‑produced, attritable unmanned aerial systems, with a clear focus on loitering munition and Counter-UAS (C-UAS) interceptor applications.

Vector‑300 has been engineered to combine advanced autonomous guidance, navigation and control (GNC) capabilities with scalability and manufacturability. Its architecture is designed to reduce technical complexity and enable agile, large‑scale production while ensuring consistent and reliable performance across high‑volume deployments.

Designed for high‑dynamic interception and terminal missions, Vector‑300 delivers strike‑to‑target precision guidance with bull’s eye accuracy. The autopilot supports the integration of AI‑based target identification and optical data directly into its autonomous GNC loops, enabling advanced engagement of both static and dynamic targets. This architecture supports real‑time trajectory adaptation during pursuit and terminal engagement phases, making Vector‑300 suitable for demanding loitering munition and C-UAS interceptor operations.

Vector‑300 is designed to operate in highly contested and GNSS‑denied environments, even under electronic warfare (EW) jamming, spoofing and meaconing. Its robust navigation core relies on advanced inertial algorithms and multisensor fusion to ensure mission continuity across all phases of operation and can be easily complemented with UAV Navigation–Grupo Oesía proprietary solutions such as the Visual Navigation System to enhance dead‑reckoning accuracy.

Building on the battlefield-proven capabilities of the Vectorautopilot family, Vector‑300 enables the full range of advanced operations already established across UAV Navigation–Grupo Oesía solutions. These include

• fully autonomous mission execution
• swarming and formation flight
• 4D trajectory management to reach targets at a predefined time
• high‑dynamic maneuvers
• manned‑unmanned teaming (MUT) operations
• many other advanced autonomous capabilities.

Its open and modular architecture is designed to ensure interoperability with third‑party platforms, payloads and sensors through seamless integration with Vector‑MCC. This architecture also enables the integration of autonomous decision‑making software, allowing platforms equipped with Vector‑300 to adapt to evolving concepts of operation and advanced autonomy requirements.

<p>The post Vector-300 autopilot designed for mass production of C-UAS interceptors first appeared on GPS World.</p>

Challenge: Capturing a massive heritage site, including every detail from courtyards and buildings down to a drawbridge and individual rivets on castle gates. 

Solution: Artec Jet, Artec Ray II, Artec Leo, Artec Twins 

Why Artec 3D? The highly maneuverable Artec Jet can be attached to a backpack and simply walked through an environment. Entire scenes can be captured from ground level in minutes, including tall structures from a range of up to 300 meters. Artec Ray II and Leo deliver higher accuracy for applications like long-term monitoring, damage assessment, and restoration. 

Odawara Castle: A gateway into Japan’s past

Odawara Castle was built more than 500 years ago, with fortifications first erected during the Kamakura period — a time famous for the emergence of the Samurai and Japan’s first Shogun. 

The site’s illustrious walls are steeped in history. Situated on a hill and surrounded by a moat, the castle has strong fortifications, so it was coveted and fought over for generations. Three sieges of Odawara took place from 1561-90 and the structure changed hands (and shape) multiple times over the next century as different leaders left their stamp on the property. 

At times, the legacy of Odawara Castle has been difficult to protect. The entire site was shaken to its foundations by multiple earthquakes from 1703-1853 and the Meiji government of the late 19th century ordered that all feudal structures be destroyed, so it was mostly torn down. 

In 1938, what remained of Odawara Castle was made a heritage site and slowly rebuilt. But over the years, it has remained a delicate piece of history in need of ongoing renovation. With this in mind, the Artec 3D support team — in Japan for a recent trade mission — opted to digitize the entire structure for future generations to enjoy using Artec Jet, Artec Ray II and Artec Leo. 

Capturing an entire castle in minutes 

When they arrived at the castle, engineers immediately understood the scale of the challenge  they were embarking on. Once one of medieval Japan’s largest fortifications, the site’s outer defensive perimeter is a whopping nine kilometers long. Odawara Castle is also a national landmark that’s open to visitors, so they didn’t have the facility all to themselves either.

This meant that speed and subtlety were critical. It would’ve been entirely possible to capture the site with a lidar, tripod-mounted Ray II, by positioning it around different areas of the fort. But this would take a prohibitive amount of time — especially when you consider that double scans are required to remove moving objects. Using Artec Jet was a lot more straightforward. 

Attaching the device to a backpack meant the castle could be scanned on foot. Walking the site, almost as if they were a tourist, was enough to capture the entire scene. Artec Jet’s remote app gave real-time feedback on scan progress, so the team didn’t leave any detail uncaptured — and compared to capture with shorter-range scanners, the time savings were enormous. 

“Artec Jet scans in a linear fashion. If it takes you two minutes to walk, it’ll take two minutes to scan — the complexity of the scene has little bearing,” explains Artec 3D scanning expert Keynan Tenenboim. “In the same time it took for Leo to scan 2-3 walls, Ray II scanned a building, and Jet digitized an entire castle. Adding in Ray II & Leo was great for areas with accessibility issues — and capturing higher detail around the walls, gate, and courtyard.” 

A Trio of Scanners for the Task

Natural environments like trees, rivers, and larger connecting spaces often offer valuable site context, but don’t need to be captured with high accuracy. Artec Jet was perfect for picking up this sort of background information, generating a continuous point cloud, and connecting the site’s more interesting features: historic walls, ornate roofs, and courtyards around the castle. 

Jet’s 300-meter range meant there was no need for ladders or scaffolding. The inner structure was captured from ground level without other visitors even noticing. Unlike Ray II, which scans from static viewpoints, Jet could also be maneuvered into difficult-to-reach areas. Both scanners are less accurate than Leo — but that’s why it’s best to combine datasets, for peak results. 

In this case, Ray II was deployed to scan the inner courtyard and gate, with Leo being used to pick up smaller details like the confined area behind the entrance. Handheld 3D scanning was also perfect for capturing a nearby medieval wall. As you can see from the scan below, fine details like tile patterns, lettering, and the wall’s internals were all captured in a single sweep. 

“This was the perfect project for demonstrating the benefits of all three scanners,” said Tenenboim. “The main castle wouldn’t be a good fit for Leo and it didn’t really fit Ray II. There was no good vantage point where we could see the facade from 100 meters away. Thanks to Jet’s range, we were able to scan from a ground level. Okay, we could’ve improved roof capture by flying Jet on a drone — but this would require more site preparation.” 

Heritage preservation with end-use potential 

Once engineers had finished scanning, they sent data back to Artec’s Luxembourg HQ via cloud sharing for processing in Artec Twins. Specifically designed to handle large datasets, Artec Twins software allows Artec Jet, Ray & Leo scans to be merged — either into a unified point cloud, or a 3D mesh that can be measured and exported to industry platforms like Autodesk Revit. 

In terms of applications, the resulting 3D point cloud would be perfect for building a virtual museum tour that allows visitors to virtually explore Odawara Castle. Regular data capture sessions would also allow site operators to monitor conditions over time. If a building’s traditional rooftop began to sag, for example, it’d be possible to carry out rapid repairs.

Deployable in seven modes: by-hand, backpack, pole, cage, robot, vehicle, or drone, Artec Jet adapts to any environment, allowing users to replace complicated multi-tool workflows. Clearly, Artec’s Odawara Castle scan is just the beginning, there are many more sites left to explore. 

See the captured dataset from this project here. 

<p>The post 3D scanning experts digitize Japan’s historic Odawara Castle first appeared on GPS World.</p>