In order to provide accurate positioning in many different environments, the output from different sensors must be combined to provide the navigation solution.
There are many different sensors available and all have their relative advantages and disadvantages. The integration of many different types of sensors ensure that the solution is:
More accurate: through exploitation of the redundancy of sensed information.
More reliable: redundancy of sensors ensure that the system is not compromised through sensor failure.
More robust: by capitalising on the advantages of each sensor, the system becomes better than the sum of its parts.
Global navigation satellite systems (GNSS) are well known and have been thoroughly tested over the years. The US global positioning system (GPS), for example, is widely used because of its high accuracy and reliability. It provides position and velocity information anywhere in the world but alone it may not be enough. Typical update rates of a GPS receiver are one per second. In highly dynamic applications, such as onboard aircraft, this is not quick enough and so other sensors are used to provide information between GPS updates. Inertial navigation systems (INSs) are a common partner to GPS receivers in navigation systems. They provide position, velocity, attitude and angular velocity information and have fast update rates (up to 400Hz). However their accuracy degrades over time. By integrating GPS and INS data, the resulting system provides accurate navigation at high update rates that don't drift over time.
There are many combinations of sensors that can be integrated:
Position. eg INSs, GNSSs, depth sensors, altitude sensors, ultra-short base-lines (USBL), map-matching.
Velocity. eg INSs, GNSSs, airspeed indicators, Doppler velocity loggers.
Acceleration. eg accelerometers.
Attitude: eg compasses, map-matching.
Angular velocity. eg gyroscopes.
Integration of these sensors is a non-trivial exercise and requires complex algorithms. Kalman filters are commonly used which take sensed information and their estimated accuracy to output precision navigation data. Our staff have developed a unique Kalman filter to provide such a function. It is configurable to take input from any type of navigation sensor including duplicated sensors (eg using multiple INSs). The filter has been thoroughly tested using simulated and real input data and has been praised by all customers who have used it. The Thales Integrated Navigation Testbed (TINT) is a simulation that uses our custom Kalman filter to drive the navigation system of each of the vehicles modelled.