UAV Navigation in depth: Dead Reckoning Operations


One of the biggest challenges facing any cutting-edge Attitude & Heading Reference System / Inertial Navigation System (AHRS/INS) is the ability to perform a mission in a degraded environment. In navigation, dead reckoning is the process of calculating one’s current position by using a previously determined one, and advancing it based upon known or estimated speeds (integrated) over elapsed time and course. Evidently each integration is subject to a cumulative error; the aim therefore is to reduce this error in order to increase precision.

An AHRS/INS uses a computer and a series of sensors, such as accelerometers, gyroscopes and occasionally magnetometers, which are used continuously to calculate: position, orientation and velocity of a navigation system or moving object without the need of external references, such as GNSS devices.

UAV Navigation develops and manufactures high-end autopilots and as such is completely committed to products which are capable of precise and reliable navigation. Navigation in degraded environments (e.g. with no GNSS input) is a major objective. There follows a description of the test results achieved in this field. Significant advances in wind estimation (which is key to good dead reckoning performance) and in other fields have been made in the UAV Navigation system which lead to greater navigation accuracy in degraded enviornments.


The equipment used to produce the following results is the VECTOR, UAV Navigation's cutting-edge autopilot. VECTOR features a highly advanced, high-end, MEMS-based AHRS/INS. It has been designed for system integration in avionics packages or other attitude sensing applications, and includes:

  • Attitude Heading & Reference System (AHRS)
  • Inertial Measurement Unit (IMU)
  • Inertial Navigation System (INS)
  • Air Data System (ADS)
  • GNSS (only for data comparison)

VECTOR has accumulated thousands of hours of flight time and has proven itself in a variety of highly dynamic environments, giving outstanding results.

Redundancy built into the software allows it to survive individual sensor failures while maintaining accurate estimates of attitude and position. Its high-performance MEMS-based IMU is calibrated and compensated over the full industrial temperature range.

In order to simulate a degraded scenario in these tests, the GNSS information was completely excluded, whilst all the other systems (see above) were functional.


In order to achieve the most realistic and representative results possible, the test was carried out in a fixed wing aircraft flying a precise Flight Plan (FP). The aircraft used is one of the fixed wing test platforms available in the UAV Navigation fleet. The platform has a 2.5m wingspan and is equipped with a single piston, two-stroke gasoline engine.

The test platform is capable of carrying all the required flight measurement devices and is able to produce repeatable results, even in highly dynamic manoeuvres. The Flight Plans and tests performed are described below.


The platform takes off from a known position and once the test altitude and location are achieved the GNSS is disabled (GNSS information is still recorded onboard, but it is not used for estimation). Two different manoeuvres are then performed which characterize the worst case dead reckoning scenarios: (1) circular hold, and (2) square pattern. Wind conditions for the tests were measured at the Ground Control Station as a constant 3m/s.

The following table explains the operations executed during the flight:

HOLD mode, circular pattern (300m radius). DEAD RECKONING

3 minutes

AUTO mode, square pattern (400x350m). DEAD RECKONING

2 minutes


The following charts show the navigation performance of the VECTOR under degraded conditions: actual position of the platform versus estimated position (derived from the integration performed by the INS).

Figure 1. Real position vs Estimated position in Dead Reckoning mode. HOLD mode.

HOLD Mode.   The results (Figure 1 above) show that in a manoeuvre lasting 3 minutes a position drift of around 100m in total is accumulated, giving a Position Drift Rate of 33 m/min.

Figure 2. Real position vs Estimated position in Dead Reckoning mode. HOLD mode.

Square Pattern.   The results (Figure 2 above) show that in a manoeuvre lasting 2 minutes a position drift of around 70m is accumulated, giving a Position Drift Rate of 35m/min.

Conclusion.   The tests prove that the UAV Navigation system extracts the very best results possible from MEMS-based sensor technology. That being said, results will vary from platform to platform and will depend on a variety of factors including: wind conditions, platform performance and quality of autopilot installation (in order to minimise vibration).