Listen to this content
GNSS unreliability in challenging environments has historically been a barrier for expanding the operational domain of autonomous driving. GNSS accuracy degrades when satellite signals reflect from surrounding structures before reaching the receiver (multipath interference), and when signals are weakened or partially blocked by obstructions.
This has prevented automotive OEMs from relying on GNSS as a trusted input for advanced driver-assistance systems (ADAS) and autonomous vehicle (AV) software stack in environments such as urban canyons, tree-lined roads, stacked roads and underpasses — the very environments in which vehicles need to navigate safely, every day.
Today, consumer expectations are pushing OEMs to extend hands-free driving beyond motorways and open roads into dense urban environments. As a result, GNSS reliability and integrity have become critical requirements for today’s ADAS.
This is the problem that FocalPoint has set out to solve, and through a strategic collaboration with STMicroelectronics, the solution is now available on automotive-qualified silicon.
Multipath interference in urban environments
A GNSS receiver calculates its position by measuring the travel time of the signals transmitted directly from satellites, a calculation that assumes each signal has traveled along a direct line-of-sight (LOS) path. Under an open sky, that assumption holds, and the receiver produces an accurate position.
In cities, however, signals frequently encounter buildings, glass facades, bridges and other structures, arriving at the receiver with a longer path. These reflected signals, or non-line-of-sight (NLOS) signals, can make the receiver overestimate the distance to the satellite, introducing errors that can reach several meters.
The “confident but wrong” problem
Multipath errors do not necessarily degrade the receiver’s confidence in its own solution. A conventional correlator cannot distinguish reflected signals from genuine LOS measurements and weights them into the navigation solution alongside good data. The receiver may continue to report a stable, high-confidence position while the actual position drifts significantly from ground truth. This creates the “confident but wrong” problem, which can be dangerous for autonomous driving, where the system is trusted to make critical decisions.
Spoofing and jamming
Spoofing and jamming compound the same underlying integrity issue. The threat is of growing concern, particularly for robotaxi operators and other L4 fleets, where a small number of high-value, geographically predictable vehicles present an attractive target for even a low-effort RF attacker.
These problems are the reason that reliability (trust in the GNSS solution) has become the decisive factor in how OEMs architect their positioning stacks. If GNSS cannot be trusted in the environments where the vehicle actually operates, that uncertainty has to be compensated for elsewhere. More weight is placed on other ADAS components such as inertial sensors, lidar, vision and HD-maps, which generally means more hardware, and a higher maintenance burden. A trustworthy GNSS input is therefore not a nice-to-have. It changes the cost structure of the whole ADAS platform.
How supercorrelation works
FocalPoint’s approach solves the problem at the receiver. Its patented Supercorrelation technology discriminates the angle of arrival of each satellite signal, allowing the receiver to focus on LOS signals and accurately measure the distance to the satellite (and therefore the position). It also suppresses the NLOS that would otherwise confuse the navigation engine.
It does this in two ways. First, Supercorrelation constructs a virtual antenna in software, which is functionally equivalent to an expensive hardware antenna component, but is implemented as a software upgrade on top of a standard automotive patch antenna. Second, it uses long coherent integration, which delivers a meaningful sensitivity improvement to the GNSS system, while using the same standard patch antenna.
The combined effect is that the receiver gains a capability it does not have in any conventional architecture: the ability to discriminate LOS from NLOS signals in real time. That discrimination is what makes the output trustworthy and gives the technology its inherent resilience to spoofing.
S-GNSS auto on ST’s Teseo devices
FocalPoint’s automotive product, S-GNSS Auto, is built on Supercorrelation and is integrated onto STMicroelectronics’ Teseo automotive GNSS devices — a platform already qualified and shipping in OEM programs. The integration is delivered as a firmware-level enhancement, so OEMs and Tier-1s can adopt it without disturbing their existing GNSS hardware path.
The combined FocalPoint and ST solution shows improvements of up to 4× in measurement accuracy and up to 3× in position accuracy in challenging multipath environments — including the city of London, Frankfurt, Tokyo and the Black Forest in Germany. In joint benchmarking, it has outperformed other commercially available competitor receivers tested in the same conditions. I am running a few minutes late; my previous meeting is running over.
Extending ADAS ODD into urban environments
The practical consequence for ADAS architects is that GNSS becomes a component they can rely on inside the urban ODD, not only on motorways and in open sky. That extends the addressable domain for hands-free features into the cities where consumer demand is high. It also strengthens the case for features that depend on an absolute, globally-referenced position such as V2X and geofencing.
It also relaxes one of the tougher mechanical trade-offs in modern vehicle design. Concealed and integrated antennas — shark-fins, roof-embedded, or hidden behind bodywork — usually come with a positioning penalty, because they receive weaker or more distorted signals than a well-placed external antenna. Supercorrelation recovers that lost performance in software, so designers can choose antenna placement on aerodynamic and styling grounds without compromising accuracy.
What’s next
Evaluation kits combining the ST Teseo platform with S-GNSS Auto are available today, allowing OEMs and Tier-1s to test the solution in their own reference routes and scenarios rather than relying on third-party benchmarks alone. Looking further ahead, FocalPoint is extending the benefits of Supercorrelation into its next-generation technology, Precise+.
Precise+ focuses on enabling high-precision carrier phase GNSS systems to keep operating in environments that have traditionally defeated them. Carrier-phase techniques such as RTK and PPP unlock centimetre-level positioning in open conditions but lose lock easily in dense urban environments. Precise+ aims to close that gap, making GNSS accurate and reliable wherever the vehicle has to drive.
