Gravity modelling is an integral part of INS. INS can only operate in situations in which the gravitational field is known, or known to be insignificant. The accurate measurement of gravity is therefore a limiting factor to the accuracy of INS. Gravity is composed of two parts, gravitation and centripetal acceleration. Gravitation is defined by Newton’s Law of Gravitation, which relates the attraction of two masses by their weight and separation. Centripetal acceleration is the rotational acceleration that occurs because of the rotation of the Earth. The combination of the two components gives the local gravity vector to which a ‘plumb bob’ would align itself if held above the Earth.

Before an INS can operate it must first determine its attitude, position and velocity. Gimballed INS require that the axes of the platform are parallel to the navigation coordinates, while for a strapdown INS the alignment involves the calculation of the initial values of the coordinate transformation from the sensor coordinates to navigation coordinates. There are four common methods employed for INS alignment, they include:

·                Optical Alignment – This takes two forms, local alignment using ground based systems such as coordinated survey marks and a theodolite or space based alignment using a star tracker, which is primarily used for alignment in space.

·                Transfer Alignment – this is achieved using velocity matching with an aligned and operating INS. This is commonly used in military vehicles to align a slave INS in a missile from the master INS in the vehicle. This requires that the onboard INS be in working order and the manoeuvres of the vehicle are suitable for the transfer of alignment.

·                GPS-aided Alignment – this uses position matching with GPS to estimate the alignment variables. It can be done without any specific movements but does take time for the navigation solution to settle to acceptable levels.

·                Gyrocompassing Alignment – this is achieved using the sensed direction of the vertical while stationary. Latitude is determined through the angle between the earth rotation vector and the horizontal. The system is incapable of calculating longitude and this has to be input into the system.

All inertial navigation systems exhibit similar qualities regardless of their configuration, gimballed or strapdown, although both systems have their nuances. The system provides information about the velocity, position and attitude, all of which degrade over time. The system does bound the effect of some errors through methods such as Schuler tuning, but others such as gyro drift cause an unavoidable continuous error growth.

The quality of the system is defined by its error growth per unit of time. The quality of the system is directly related to the quality of the components. There are many applications for which inertial navigation systems are used, both short and long duration. Short term applications generally do not require sustained performance and consequently use systems that minimise the error growth for a short period after which is not considered. In these situations the quality of the components is not paramount and cheaper alternatives can be used. Longer duration applications require the use of precise instrumentation but continuos error growth is still unavoidable. The point at which the solution becomes unusable is determined by the requirements of the application and the instrumentation. The ability to determine this point is only capable through the known or estimated error growth of the system.

The system is also influenced by gravity, the system has to apply a value of gravity for each calculation and the inability to perfectly model the effect of gravity causes errors that continually propagate in the vertical channel. This can transfer to horizontal measurements however the effect is bounded in these channels, unlike the vertical channel.

The system is almost self-reliant, but it is dependent upon the input of data to determine its initial position, even through gyrocompassing the system cannot determine longitude. The system does provide a continuous data stream once operational and achieves this at a high output rate. The system is autonomous once operational and therefore is incapable of being jammed or detected in most circumstances.