Reliable navigation systems need multiple points of reference to ensure accurate positioning. GNSS can provide highly accurate location data, but signals can be blocked, degraded, jammed, or spoofed. For military platforms, autonomous systems, and other applications operating in contested environments, this creates a requirement for GNSS-denied navigation (often called GPS-denied navigation): positioning that continues when satellite signals cannot be trusted.
Inertial navigation provides an important part of platform resilience and is increasingly used on a number of platforms. However, an Inertial Navigation System (INS) will accumulate positional error over time without an external reference, creating challenges for users. This is known as INS drift.
This can be addressed through multi-sensor fusion navigation, which combines inertial measurement with other sources of navigation data, typically brought together by a Kalman filter.. This means that rather than relying on one sensor, the navigation system builds its solution from several systems to validate its measurements.
Leveraging a fused sensor network
For an entry-level non-GNSS solution, military and commercial platforms can use an Inertial Measurement Unit (IMU) for positional data, a system that leverages gyroscopes and accelerometers to measure angular rate and linear acceleration. For additional accuracy, users can employ an INS, which processes this data to estimate a more detailed position, velocity and attitude.
What causes INS drift?
As inertial navigation systems do not depend on an external signal, they can continue to operate when global navigation satellite systems (GNSSs) are unavailable. Their principal limitation is inertial navigation drift, which are small measurement errors that accumulate as the system calculates movement over time, gradually reducing positional accuracy.
Employing external aids to validate measurements can limit these errors.
During normal operation, GNSS commonly provides this reference. In GNSS-denied conditions, other sensors can provide critical information about the movement of a platform over space and time.
Air: air data computers
For airborne platforms, including crewed aircraft or UAV, an Air Data Computer (ADC) can provide information including airspeed and pressure altitude. These measurements then support the INS in correcting velocity and altitude estimates.
Maritime: DVLs and Doppler logs
In marine applications, technology such as a Doppler Velocity Log (DVL) provides an independent measurement of velocity for submersibles based on their relative speed and direction from the seabed, while a Doppler Log is used for surface vessels and measures position relative to the water or seabed.
Land: odometers and wheel sensors
For ground vehicles, wheel sensors, encoders, or odometers can provide additional information about vehicle movement.
Other technologies, including visual odometry and RF ranging, can introduce further independent measurements. See more on our Visual INS solutions.
The aim is not necessarily to find a direct replacement for GNSS as it is still a key position, navigation and timing source. Instead, different sensors contribute information that helps build a more reliable navigation solution and aid situational awareness.
How a Kalman filter corrects INS drift
Combining several sensors can create a technical challenge as their measurements often have different characteristics, update rates and levels of uncertainty. To help provide a framework for bringing this information together, developers such as Inertial Labs rely on Kalman filtering.
At a simplified level, the process continuously alternates between prediction and correction. The navigation system first predicts its new state using its previous position, velocity and attitude combined with the latest inertial measurements. It also estimates the uncertainty associated with that prediction.
When another measurement becomes available, such as GNSS position, DVL velocity or air data, the filter compares this with the predicted state. It then adjusts the navigation solution according to the estimated reliability of the available information.
This makes Kalman filter sensor fusion more sophisticated than simply averaging multiple measurements. The system is continually estimating the most probable navigation state from imperfect information.
GPS-denied navigation in practice: a UAV example
The value of this architecture becomes particularly clear when GNSS coverage disappears, either through natural or man-made reasons.
Consider a UAV operating in a GNSS-denied environment: its IMU continues measuring acceleration and rotation, allowing the INS to propagate the navigation solution, while at the same time, an ADC may continue providing airspeed and pressure altitude.
Depending on the platform, further sensor information could come from visual odometry, RF ranging or other alternative navigation sources.
Each source addresses a different part of the navigation problem. Airspeed can support velocity estimation, pressure provides an independent altitude reference, while visual odometry can estimate movement by analysing changes between successive images.
By combining this information, the navigation system can reduce the rate at which errors accumulate during GNSS outages.
Countering GNSS jamming and spoofing
GNSS jamming and spoofing are increasingly common in conflict zones where dedicated electronic warfare assets are deployed, which means GNSS may be not just missing but actively misleading.
Jamming can prevent a receiver from obtaining usable satellite signals, while spoofing can be more difficult as false GNSS information being fed to a platform may initially appear legitimate.
Ensuring that a system has a diverse range of sensor data allows for greater and more trusted assessment of positional information, which is mission critical for end users.
If GNSS-derived movement is inconsistent with inertial data and other measurements, the system has more information available than a platform relying predominantly on satellite navigation.
This layered approach is central to Assured Positioning, Navigation and Timing (APNT), also known as assured PNT . And of course, the objective is not to eliminate GNSS, but to prevent the loss or degradation of what could potentially be a single point of failure.
Matching aiding sensors to land, air and maritime platforms
There is, however, no universal sensor-fusion architecture. Open-architecture platforms still rely on a data interchange with sensors, which need a translation layer.
An unmanned ground vehicle has different requirements from a fixed-wing UAV or autonomous surface vessel for movement, control, and navigation. The most appropriate aiding sensors therefore depend on the platform, environment and mission.
Inertial Labs’ navigation systems support integration with external sensors across land, air and maritime applications, including DVLs, odometers, wheel sensors, encoders and air data sources. Its wider APNT technologies also extend into areas such as visual navigation and alternative RF-based positioning.
The INS, however, remains at the centre of this architecture, providing continuous high-rate navigation between external updates.
Building resilient navigation with assured PNT (APNT)
The development of APNT is changing the way navigation systems are designed. Rather than seeking the technologies that could replace GPS, it is more important to find how multiple independent technologies can work together – maximising the investment in sensor deployment.
High-quality inertial sensors provide continuity, which are supported by external sensors that provide data to minimise inertial drift. Kalman filtering provides the framework for combining those measurements into a coherent estimate of position, velocity, and attitude.
For autonomous and defence platforms expected to operate through jamming, spoofing and intermittent GNSS availability, the combination of inertial navigation, complementary assisting sensors, and intelligent sensor fusion provides the foundation for reliable GPS-denied navigation, keeping INS drift in check whether GNSS is lost, jammed or spoofed. .
Frequently asked questions
What is GNSS-denied navigation?
GNSS-denied navigation is the ability to keep determining position, velocity and attitude when satellite signals are blocked, jammed or spoofed. It usually relies on an inertial navigation system aided by other sensors, such as air data, DVLs, odometers or visual odometry.
What causes INS drift?
INS drift is caused by small errors in gyroscope and accelerometer measurements. Because an INS calculates position by continuously integrating these measurements, the errors build up over time and gradually reduce accuracy unless an external reference corrects them.
How does a Kalman filter fuse INS and sensor data?
A Kalman filter predicts the platform’s next position, velocity and attitude from inertial data. When another sensor provides a measurement, the filter compares it with that prediction and corrects the solution, weighting each input by how reliable it is.
What is the difference between GNSS jamming and spoofing?
Jamming overwhelms satellite signals so the receiver can’t get a usable fix. Spoofing broadcasts false signals that appear legitimate, so the receiver calculates an incorrect position. Spoofing is often harder to detect.
What is APNT?
APNT, or assured PNT, stands for Assured Positioning, Navigation and Timing. It’s a layered approach that combines GNSS, inertial navigation and alternative sensors so that a platform doesn’t depend on any single source of navigation data.
