An inertial navigation system is essentially a set of inertial measurement units with a timekeeping component, usually paired with pilotage or cartography data. It is a mechanism designed to keep track of its current instantaneous position in at least one (but usually two or three dimensions) given only a known starting point and vector, using no external reference sources.

MIlitaristic jargon aside, what does that all mean? It means a black box that, once initialized, can from that point on tell you where it is in relation to your start point without having to see outside or talk to anything else. This is extremely useful for vessels which do not want to or cannot communicate with outside references while underway - ships at sea, submarines, aircraft in bad weather or flying under stealth, guided weapons which just need to know how to get to one place without being jammed by outside influences, etc.

### How does it work?

Without going into the math, the concept is fairly simple. Imagine that the inertial navigation system (INS hereafter) is a black box, sitting at rest. Now, it contains IMUs, which will tell it whenever it is accelerated in any direction, and a good chronometer that will tell it precisely for how long it has been accelerated. Knowing those two bits of data, it is designed to do the basic physics and figure out its resulting velocity - and, as time goes on, its positional change as a result of that velocity and any changes in velocity that occur.

Although it's not required by the definition, most INS systems also have a means of relating their current position to a set of known map or chart data, presenting the user with a ship's position. Since the INS is mounted within the ship, it can be assumed that the ship goes with it.

### What are their limitations?

Their advantages have been enumerated above - entirely self-contained position data, resistance to interference, no infrastructure required. Their main limitation is that of accuracy. No physical system is perfect, and since an INS utilizes integrals, any errors in the measurements or calculations begin to 'build up' geometrically in the reported position. As a result, an INS must be periodically 'reset' to a known position to eliminate these built-up errors. For most applications this isn't a problem - aircraft flights don't last that long, and ships can cross-check the results against celestial navigation or artificial navigation aids such as LORAN or GPS, these days. Submarines, however, have to rely on INS for longer periods, and hence need more precise units and, more importantly, a very precise knowledge of how accurate their systems are so that they know when to stop trusting them.

In the early days, INS systems were also fragile, relying as they did on gyroscopes not only for measurement of acceleration but on gyroscopic isolation from any rotation or vibration of the surrounding craft. This was accomplished through the use of a 'stable platform' - a table which was kept level through the use of an additional set of gyroscopes tasked only with keeping its surface isolated from rotation. The problem, of course, was was that this introduced yet another source of error into the INS measurements. Eventually, with the advent of optoelectronic systems such as the ring laser gyro and the fiber-optic gyroscope, the fragility of the system lessened, and the advent of computers powerful enough to implement strapdown guidance meant the stable platform itself could be dispensed with.

### Where did they come from?

The first U.S. military INS was called the SINS, (Ship's INS) and was finally approved and built for the Polaris Fleet Ballistic Missile System submarines, which were the first U.S. submarines whose mission called for them to remain submerged and unable to perform celestial navigation for days or weeks at a time. Precursors of the SINS had been around for years, however; the gyrocompass used by submarines and ships throughout most of the twentieth century would eventually lead to the gyroscopic IMU. Those interested are strongly encouraged to read the excellent book Inventing Accuracy: A Historical Sociology of Nuclear Missile Guidance by Donald Mackenzie (MIT Press: Cambridge, 1993). The first 'deployed' SINS went to sea with the first Fleet Ballistic Missile boat, SSBN-598 George Washington, which was launched on June 9, 1959.

### Where are they used now?

Everywhere! INS are used in aircraft of all sorts, in ships, submarines, trains, satellites, missiles; even in cars, trucks, and boats. There are models you can carry with you, for when GPS isn't available, although their cost usually restricts their availability to governments and the very well-heeled. INS systems can and even have been put in the noses of artillery shells to tell them where to go.