Multisensor Navigation

Multisensor navigation is the process of estimating the navigation variables of position, velocity, and attitude from a sequence of measurements from more than one navigation sensor. There are essentially two broad categories for sen­sors used in avionics suites for navigation and related functions: dead-reckoning sensors and positioning sensors.

Dead-reckoning sensors provide a measure of acceleration or velocity with respect to an Earth-referenced coordinate system, consequently requiring inte­gration with respect to time to provide vehicle position with respect to the Earth. Examples of these types of sensors are inertial systems, Doppler radars, and air-data sensors. The latter two require an attitude and heading reference (AHRS) or an inertial system (INS) to provide the required angular orientation with respect to the Earth.

Positioning sensors provide a position measurement that can be related to Earth-referenced coordinates. Examples of these sensors are radio systems such as the terrestrial-based Loran and the satellite-based Global Positioning System (GPS) which provide the position of the antenna in geodetic coordinates. A star-tracker can also be used for fixing position when its orien­tation with respect to the Earth is determinable through some means such as an AHRS or INS.

In general, avionics system users require a variety of information on the state of the air vehicle depending on the com­plexity of the application—aerospace plane to auto-gyro—and accordingly will include a variety of complementary sensors in their equipment suite. The infor­mation desired can include the following:

· Position and velocity in geodetic coordinates—east, north, and up which allow determination of ground speed and track angle.

· Orientation with respect to the Earth—pitch, roll, and yaw or heading angles.

· Linear and angular acceleration and rate in body coordinates for vehicle control purposes.

· Vehicle state relative to the air mass including orientation—angle of attack, sideslip, and airspeed, again for vehicle control purposes.

Clearly, the entire list of state variables for the vehicle cannot be provided by any one sensor at normally desired levels of accuracy and dependability. The deficiencies that exist in the individual sensors employed in a multisensor avionics suite include at least one of the following characteristics:

• Increase in error of the navigation variables as a function of time or distance traveled. This is a characteristic of all dead-reckoning sensors that eventually accumulate unbounded errors. Examples are inertial and Doppler radar navigation systems.
• High noise level or low bandwidth in any derivative variable. This is a characteristic of radio sensors that require differentiation with respect to time of the basic measured variable to obtain rate or acceleration. For example, with Loran, position must be differentiated to obtain velocity. For Doppler radar, differentiation of speed can yield acceleration, once the result is transformed to an Earth-referenced coordinate system using attitude from an AHRS or inertial system.
• Reliance on sensors requiring off-board components to accomplish the required functions. This includes radio sensors using terrestrial or satellite-based assets whose access can be denied, due to intentional (jamming) or nonintentional interference of the propagated signal from the off-board assets, or failure of the transmitters due to other causes.

To overcome individual sensor deficiencies, system designers have sought combinations of avionic sensors. These multisensor systems are designed to provide reliable, dynamically accurate measurements of the air vehicle's state for all specification-required flight conditions.

The most typical solution has been to use an inertial system in conjunction with an appropriate set of complementary sensors that arrest the random, time-increasing error in velocity, position, and orientation resulting from integration of the fundamental high-bandwidth inertial measurements of acceleration (force corrected for the a priori known effect of gravity) from the accelerometers and angular change from the gyroscopes.

Two fundamental error sources affect the error behavior of an inertial system. These are the errors in the measurements of force made by the accelerometers and the errors in the measurement of angular change in orientation with respect to inertial space made by the gyroscopes. These measurements are then compensated for the force of gravity with a mathematical model, such that vehicle acceleration with respect to inertial space is obtained. The resulting variable, after correction for Coriolis accelera­tion, is then integrated once into velocity and a second time into position change with respect to the Earth. Additionally, the gyroscopic measurements of angular change with respect to inertial space are modified using the system computed velocity and the Earth's rotation rate vector to reflect the rotation of the local vertical due to earth rotation and the vehicle change in position as it travels over the surface of the Earth. In this manner the orientation of the accelerometer axes relative to the Earth at the present position of the vehicle is continuously computed. The result is that there are three sources of change in orientation error of the accelerometers with respect to an Earth-fixed reference coordinate frame: (1) integrated gyro drift rate; (2) integrated error in system computed velocity which results from error in the measurement of acceleration—due to accelerometer measurement errors, imperfect knowledge of the local force of gravity and the current error in the knowledge of orientation of the accelerome­ter sensing axes which causes a misresolution of any accelerometer force mea­surement including that of the gravity vector; and (3) error in the orientation of the navigation coordinate axes which changes as they rotate with respect to inertial space.