Elektron Systems, Inc.
|
|
| Fiber Optic Gyro Based Northfinders | |
|
Elektron Systems has been involved in three separate developments of all fiber optic gyro based
northfinders. The latest of these programs is an internally funded effort to develop a small size and low cost (< $10,000 single unit price) northfinder for commerical and military (COTS) applications. The particular northfinder implementation being developed requires that the northfinder be stationary during the acquisition process. This is unlike inertial navigation units and gyrocompasses that can find and maintain a north reference on a moving body, but that also are quite a bit more expensive. It is also unlike using a long baseline with GPS or a multiple antenna GPS azimuth determination implementation, both of which are subject to signal blockage and jamming. The northfinder, by measuring the Earth's rotation directly, is inherently un-blockable and un-jammable. And, unlike a magnetic compass measurement, it is not subject to external influnences and does not point to magnetic north. A fiber gyro northfinder provides an azimuth reference relative to true north. A northfinder can provide an alignment in heavly wooded areas, mountainous areas, in cities, in buildings, underground in tunnels and mines or drilled shafts and under water, as long as the platform it is mounted on can be made stationary during the acquisition process. It is even possible that the gyro used for northfinding while stationary, can also be used to maintain the initial azimuth reference into and during a period of motion. The Method Most of the fundamental components of this type of northfinder can be seen in ESI's breadboard shown in the image below. A rotation mount for the gyro is provided with some sort of motor to turn it. In this case a stepper motor is used to turn a large brass gear serving as the rotation platform. On the platform the gyro is mounted with it's sensitive axis horizontal (this image does not have a gyro in place on the vertical mounting bracket). A laser pointer and three axis tilt sensor are mounted to the back side of the gyro mounting bracket. Under the brass gear are found the rotation bearing and slip ring assembly. The slip ring is used to pass signals from hardware on the rotation platform to the stationary components. A means of measuring the position of the rotation platform is provided by an opto-interrupter, and the stepper motor's control circuit maintain the position count from the initial zero reference. The breadboard uses a small microcontroller to place the opto-interrupter, stepper motor drive and interface to the gyro under control of a PC via RS-232 interface. |
|
|
|
|
Custom software running on the PC turn the breadboard and its gyro into a northfinding system. The software not only provides the houskeeping functions of acquiring gyro data, running the rotation platform and reading the tilt sensors. It also executed the algorithms to reduce the acquired data into azimuth measurements and display and save the results. The process runs in the following manner: The platform is rotated until a zero reference is obtained from the opto- interrupter. From this point on any rotation commands to the stepper motor are counted to maintain a platform position value (Measurements with the laser pointer on this breadboard have shown that even this relatively simple implementation can maintain a position accuracy in excess of one part in 10,000.) The platform is then rotated to an initial position and gyro data acquired for a pre-determined period of time. Once completed, the gyro is then rotated to a new position and more data acquired. This process is repeated for a number of positions. Once all the data is acquired it is processed to determine the direction to true north. Each acquisition of gyro data represent the rate of rotation of the Earth in the plane of the rotation platform for the direction the gyro is pointing. These individual directions and rotation magnitudes can be combined to compute the direction in the rotation platform plane of the Earth's rotation vector. A measurement of the tilt of this plane can then be used to project from the platform plane to a local horizontal plane the measurement. In the local horizontal plane the measurement represents true north. Of course this is a simplification of the process. In reality, a number of additional steps are taken to record the data and reduce it to an azimuth reference to minimize artifacts in the process and in the gyro measurements. ESI has developed several numerical algorithms to reduce measurement data to azimuth reference information. These algorithms also include functions to increase the measurement accuracy by minimizing error sources in the sensor data. The Performance There are various limits to the accuracy of a northfinder. One is the latitude at which it is operating. Northfinding at the equator is the easiest, moving north (or south) makes the process more difficult. This is a result of measuring the rotation in a plane that is mostly locally horizontal. As the northfinder is moved away from the equator the magnitude of the measured rotation becomes smaller and smaller, decreasing the signal to noise ratio of the measurement. Another significant limit is the noise and bias drift of the gyro being used for the measurement. Finally, the magnitude of the Earth's rotation needs to be considered, 15 degrees per hour, maximum, or 0.00416 degrees per second. Typical northfinding times are measured in minutes. The following graph shows actual measurement accuracy as a function of measurement time for the breadboard illustrated above with a low cost gyro of 75 meter coil length mounted. As can be seen the initial measurement accuracy increases rapidly with time and then starts to level out due to noise and bias drift effects. Actually the performance is quite good considering the short coil length of the gyro. Two degrees RMS is at least as good as a compass, in the field, and this is to true north in two to three minutes. (Beware of the compass people when they talk about accuracy. Yes, a compass can have an accuracy of 0.1 degrees, but that is its angular measurement accuracy. In reality, numerious effects conspire to give a variance and deviation to the Earth's magnetic field at any particular point so that the accuracy to magnetic north in general will be on the order or one to two degrees and can be much worse.) Lengthing the fiber in the gyro's coil is one way to increase performance by improving scale factor. Increasing optical power into the fiber is another improvement that decreases noise. |
|
|
|
|
In applications where an initial azimuth reference is needed fast a northfinder can be configured to output an early estimate. Then allowing the northfinder to continue to run, that estimate can be further refined. In the graph it can be seen that the RMS error reduced to a limiting value around 0.75 degrees in 45 minutes of operation using this particular gyro. The Development Elektron Systems has a goal of a northfinder in a cubic enclosure of less then 6 inches on a side, with an average power usage of 3 watts during acquisition, and accuracy of one degree in one minute, over the temperature range of -40degC to +70degC for under $10,000(USD) in single quantity. To achieve this goal Elektron Systems is developing a low cost custom fiber optic rate sensor specifically configured for northfinding. This sensor will be combined with an advanced mechanical positioner and a DSP processor for both sensor operation and north computations. Elektron Systems is also developing advanced demodulation algorithms for fiber rate sensors that will significantly reduce or eliminate errors brought about by mechanical vibrations. This is a problem that has plagued previous fiber gyro northfinder implementations. Should you have interest or need of this technology please contact Mr. Joseph Malek for a progress update. |
|
| Elektron Systems, Inc. | |
| 4027 Alabama Ave. NE | |
| St. Petersburg, FL 33703, U.S.A. | |
| - | |
| Telephone: 727.525.4498 | |
| Email: joe.malek@eleksys.com | |
| - | |
| Last updated: January 16, 2006 | |