Industries / Transport and vehicles / Aerospace and defense / Space and satellites
Satellite navigation
Satellite navigation is broadcasting timed radio signals from satellites so that a receiver anywhere can work out its position and the exact time; GNSS, for Global Navigation Satellite Systems, is the general name for GPS and its counterparts.
6
Milestones achieved, from 1964 to 2020.
2
To watch for: announced, not yet confirmed outside.
2
Open deliverables. Nobody has shipped them.
The road
AchievedWatch forOpen
Achieved
1964
Navy satellites guide ships
After Sputnik's launch in 1957, scientists at the Johns Hopkins Applied Physics Laboratory (APL) found that the Doppler shift of its radio signal, the change in pitch as it passed overhead, could predict when it would be in view. Turned around, a known orbit and a measured Doppler shift give the listener's position. APL built the Transit system on that idea for the US Navy's missile submarines. It became operational in 1964 and was released to civilian ships of all nations in 1967; by 1990 more than 80,000 receivers had been made, with sets for pleasure boats starting at about $1,000.
Credit
Johns Hopkins APL, for the US Navy
1978
First GPS satellite flies
The Global Positioning System (GPS) finds position from the travel times of signals from several satellites at once, each carrying an atomic clock. Bradford Parkinson, who ran the US Air Force program, called a Pentagon meeting over Labor Day weekend in September 1973 that settled the design; the program was approved on 14 December 1973, and the first GPS satellite launched in February 1978, by the account of Parkinson and Stephen Powers.
Credit
Bradford Parkinson and the US Air Force; Ivan Getting of the Aerospace Corporation and Roger Easton of the Naval Research Laboratory for earlier work
2016
Wrong time from healthy satellites
On 26 January 2016 GPS satellites that reported themselves healthy broadcast a wrong value for converting GPS time to Coordinated Universal Time (UTC), the world's civil time. The US Air Force said a fault in its ground software, set off when the oldest satellite, SVN 23, was removed from service, put the UTC signal off by 13 microseconds, beyond the design limit, while positions stayed correct. Meinberg, a German clock maker, saw time jumps of about 13 microseconds on all its GPS receivers from about 02:00 to 12:00 UTC, and reported effects on precise timing receivers in Europe, North America and Asia.
LessonNetworks that take their time from GPS need a check that does not come from the same system.
Credit
US Air Force 2nd Space Operations Squadron (operator); Meinberg (outside analysis)
2016
Galileo opens, clocks fail
The European Commission declared initial services for the European Union's Galileo system on 15 December 2016, with 18 satellites in orbit. In January 2017 ESA reported clock faults on five of the 18. Each Galileo satellite carries four atomic clocks, two rubidium and two hydrogen maser; three rubidium clocks had failed, and six maser clocks had failed over the previous two years. ESA said services were not affected, linked the rubidium failures to probable short circuits, possibly from a test procedure on the ground, and said clocks not yet launched needed refurbishing.
LessonFlaws in a part can surface only after months in orbit, across several satellites at once.
Credit
European Commission (owner); ESA
2018
Two-frequency fixes in phones
Satellite signals are slowed by the ionosphere, a charged layer of the upper atmosphere, and in cities they bounce off buildings. A receiver that tracks two frequencies can remove much of the first error, and the newer L5/E5 signals are easier to tell from reflections, according to the EU Agency for the Space Programme (EUSPA). On 31 May 2018 Xiaomi launched the Mi 8 with Broadcom's BCM47755 chip, which EUSPA calls the world's first dual-frequency GNSS smartphone. Robustelli and co-authors computed positions from the phone's raw measurements in a study published in January 2019, and found weaker signals and more reflection error than in a survey-grade receiver.
Credit
Broadcom (chip); Xiaomi (phone)
2020
China completes BeiDou
China declared its BeiDou-3 navigation system commissioned on 31 July 2020, which China Daily called the formal opening of the system to users worldwide.
Credit
China Satellite Navigation Office and the Chinese government
Watch for
Navigation from low orbit
Xona Space Systems is building Pulsar, a fleet of navigation satellites in low Earth orbit, about 500 km up, far closer than GPS. Its test satellite, Pulsar-0, launched in 2025, mapped GPS jamming from orbit along a band from France to the Pakistan border; in the worst areas the satellite's GPS signal-quality reading fell from a normal 40 dB to as little as 10 dB, by Xona's account. In August 2026 Xona said the US Federal Communications Commission (FCC) had approved broadcasts from its full constellation, that more than a dozen commercial receivers had tracked more than 350 passes, and that its signals are up to 100 times stronger than GPS; the first six production satellites were slated to launch in October.
Achieved whenpaying users rely on the signals and an outside tester publishes position and time accuracy with and without GPS.
Claimed by
Xona Space Systems, June and August 2026
Signals that prove their source
Spoofing is broadcasting fake satellite signals so that a receiver shows a false position or time. Galileo's Open Service Navigation Message Authentication (OSNMA) adds a digital signature to the free signal, so a receiver can check that the navigation data came from Galileo. The European Commission declared initial service on 24 July 2025, and it has been mandatory since December 2025 in Europe's smart tachographs, the recorders of truck drivers' hours. In February 2026 ESA reported that receivers using it told authentic messages from modified ones at Norway's Jammertest event; those are ESA's own tests.
Achieved whenan outside lab or regulator publishes field results showing receivers on sale rejecting spoofed Galileo data, with a count of misses.
Claimed by
European Commission and ESA, July 2025 and February 2026
Open
01
Phone baselines for jamming maps
A public table, for the 20 most common Android phone models, of how each model's automatic gain control (AGC) reading behaves when nothing is wrong. AGC is the receiver's own volume control; it turns down when extra radio noise arrives, so a drop can reveal jamming. The table would let a crowd-sourced, ground-level jamming map tell real interference from one model's normal behavior.
The briefCredit · unclaimed02
Holdover table for timing modules
An independent, published test of how far the clocks in common GNSS timing modules drift once they lose satellite signals, a period called holdover. Each module's antenna is disconnected and its time error is measured against a reference that stays locked, at 1, 4 and 24 hours.
The briefCredit · unclaimed
Open 01 · the brief
Phone baselines for jamming maps
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A public table, for the 20 most common Android phone models, of how each model's automatic gain control (AGC) reading behaves when nothing is wrong. AGC is the receiver's own volume control; it turns down when extra radio noise arrives, so a drop can reveal jamming. The table would let a crowd-sourced, ground-level jamming map tell real interference from one model's normal behavior.
Why it is still open
Xona says its jamming map from orbit may not show where jamming is worst for people on the ground. The European Commission's Joint Research Centre built a phone app and server that collect AGC readings and reported in 2026 that small AGC changes cannot always be linked to interference and that phone models differ widely, which has to be handled before crowd-sourcing works. A February 2026 study by Glädje Karl Olsson and co-authors located a jammer with three phone models and found one model's AGC "not usable, reporting a value of zero most of the time." A search found no published baseline across many models.
Built with what exists
Android's GNSS interface, which reports AGC, and Google's GnssLogger app, which records it; several phones of each model placed side by side; Python for analysis. Nothing is transmitted; the phones only listen.
First step
Record GnssLogger files from five phone models placed side by side for seven days and plot each model's AGC spread.
Done looks like
For at least 20 phone models, each model's normal AGC range measured over at least a week at three sites, published with code; and, on logs gathered in a region where interference is reported, a detector using the table that flags interference periods with a stated false-alarm rate per phone-day.
Open 02 · the brief
Holdover table for timing modules
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
An independent, published test of how far the clocks in common GNSS timing modules drift once they lose satellite signals, a period called holdover. Each module's antenna is disconnected and its time error is measured against a reference that stays locked, at 1, 4 and 24 hours.
Why it is still open
Phone networks, broadcasters and power grids take their time from GNSS, and the 2016 GPS fault put precise timing receivers off by about 13 microseconds for hours. Published work models holdover in theory (Demetrios Matsakis and Mathew Slavney, 2025) or tests one timing receiver against interference (the EU's STRIKE3 project, 2019); a search of published papers found no side-by-side measured table of the modules on sale. The timing company TimeBeat published a holdover test method in April 2026 and said it expected to publish measurements of its own products in mid-2026; in November 2026 its page still showed only datasheet figures.
Built with what exists
Timing modules from several makers; a time-interval counter; a second GNSS-disciplined clock left connected as the reference; Python with the allantools library for clock-stability analysis. Unplugging an antenna is lawful and needs no license.
First step
Run two timing modules locked for a day, disconnect one antenna, and log its time difference from the other every second for 24 hours.
Done looks like
At least 10 modules, each tested three times, with time error at 1, 4 and 24 hours after the antenna is pulled, room temperature logged, and each result set beside the maker's stated holdover figure, with all raw data public.