Communications: directive radio wave systems and devices (e.g. – Directive – Position indicating
Reexamination Certificate
2000-05-03
2002-03-19
Issing, Gregory C. (Department: 3662)
Communications: directive radio wave systems and devices (e.g.,
Directive
Position indicating
C342S195000, C342S059000
Reexamination Certificate
active
06359586
ABSTRACT:
TECHNICAL FIELD OF THE INVENTION
The present invention relates in general to tracking of targets by means of measurements from various sensors and in particular to automatic sensor alignment.
PRIOR ART
Target tracking is the process by which the location and motion of objects like aircraft, ships, missiles, etc., are established and presented to users, for example, situated at a command and control centre. The detection of the targets is accomplished by one or more sensors. The most widely used sensor for targets in the air is the radar. However, radars are not the universal solution since they are impaired by serious inherent drawbacks. At the state of the art, their accuracy is limited. Moreover, they can't be used under water. Additionally, by radiating energy, radars reveal their locations and form easy targets for enemy forces. Depending on the particular application, one therefore uses other types of sensors instead of or in complement of radars. Examples of such sensors, which, however, should not be considered as a complete set of possibilities, are Global Navigation Satellite System (GNSS) and its predecessors GPS and GLONASS, Electronic Support Measures (ESM) equipment, Electro-Optical (EO) sensors and their subclass Infrared Search and Track (IRST), Jam strobe detectors (ECM), air pressure based altitude measuring equipment downlinked via ‘Secondary Radars’ (Mode C), as well as active and passive sonars. Out of these sensors, only certain radars (3D) and the satellite based navigation data provide full three-dimensional determination of the target location. ESM, EO, IRST, ECM, and passive sonars only give the direction to the target. Such measurements are referred to as “strobes”.
The sensors produce measurements at regular or irregular time intervals. A measurement is a piece of information including e.g. range and azimuth of the target. A tracker is a device, which will create a track file for each detected target. A track is a set of usually filtered data which are associated with a certain target. The tracker updates the track data with incoming measurements in such a way that measurement noise is reduced by filtering, and speed and heading is computed. The track is given a label and is presented to the operators. The general art of tracking is well known and can be found at many different places, e.g. in S.S. Blackman “Multiple-Target Tracking with Radar Applications”, Artech House, Inc, MA, USA 1986, in particular pages 1-17 and 357-393.
There are many reasons to use multiple sensors for target tracking. Evidently, the combination of passive sensors at different locations makes it possible to obtain the distance to the target. By using several radars, gaps in coverage of individual radars may be eliminated. Moreover, satellite data will improve accuracy, if available. ECM gives information even if a jammer on the target makes the radar data useless, and so forth. In general, several sensors contribute to more frequent updates, better accuracy, higher signal to noise ratio, and more reliable tracking.
There is, however, a serious problem in combining data from several sensors: alignment errors, also known as systematic errors, bias errors, or registration errors. The alignment errors may give the impression that the number of targets is larger than the actual number because the measurements are displaced by some, often very significant, amount. Even if there is only one track per target, accuracy will clearly be impaired. Therefore, estimation and compensation of the bias errors are necessary for successful tracking.
To illustrate these problems, a simple example will be discussed.
FIG. 1
shows an example of a tracking situation. The tracking system have measurements available from two 3D sensors, D
1
3
and D
2
3
, respectively, providing range, azimuth as well as height information, and one passive sensor D
1
, only providing azimuth angle measurements. One true target X is present in the area covered by the system. The sensors exhibit some systematic errors. The local coordinate system of sensor D
1
3
is misaligned with respect to the common coordinate system with a certain angle &agr;. The sensor D
2
3
has an error in the range measement, which gives rise to systematic errors in the range values. In the situation in
FIG. 1
, sensor D
1
3
detects a target in the direction of a real azimuth angle A
R1
3
, but due to the misalignment, the apparent azimuth angle A
A1
3
is shifted an angle &agr;, so the apparent location for the detected target is M
1
3
. The range measuring error of sensor D
2
3
causes the apparent target position as measured by sensor D
2
3
to be M
2
3
, offset a distance r along the azimuth angle from the true position. Sensor D
1
provides a true azimuth angle measurement. In this situation, it is probable that a target tracking system will interpret the situation as if two targets are present, at M
1
3
and M
2
3
. The M
1
3
position is reasonably consistent with measurements from both D
1
3
and D
1
, while the M
2
3
position is far away. From this example it is obvious that there is a need for some alignment procedure.
It turns out that the bias estimation or automatic alignment is quite a difficult task, and it is further complicated by drifts in electronic and/or mechanical systems, which call for recurrent recalibration. The methods existing so far have various severe limitations. Some approaches according to prior art are described below.
A first approach concerns estimating and compensating for the deviations in x, y and z for each single track separately. The patent DE 3,132,009 gives an example of this approach. The drawback is that the estimation process is initiated for each track, and it may take considerable time before the estimates are good enough for accurate tracking.
Most other methods aim at modelling each sensor with a set of bias parameters. In this document, the term “bias parameters” is used for parameters, each of which represents a characteristic of the sensor like a location error, a north alignment error, or a range offset etc. Another term used in the literature, basically denoting the same measures, is “registration errors”. Most methods according to the state of the art handle a very limited set of sensor types and a very limited set of parameters.
A second approach of automatic alignment uses the above described principle which is applied by using reference targets with exactly known positions. By comparing a set of measured positions with a set of true positions one can adapt the bias parameters so that the deviations are minimised. There are, however, severe disadvantages. The positions of the reference targets have to be measured very accurately. The position of the reference target should optimally also be situated somewhere in the volume, where normal targets appear, which often may be over sea or even over international or other nations territories. Fixed reference targets with well defined positions are also normally restricted to ground level or close to ground level, which makes calibrations of altitude and inclination measurements difficult to perform. Moreover, reference targets are also easily destroyed or manipulated with during hostile situations.
In the absence of reference targets, a third approach involves designation of one of the sensors as a reference sensor, and alignment of the others with respect to the reference. One example of such a system is found in the proceedings of the SPIE—The International Society for Optical Engineering, Vol. 2235, 1994, in an article in the name of R. Helmick et al, with the title “One-step fixed-lag IMM smoothing for alignment of asynchronous sensors”, pages 507-518. Here a method for compensating a number of 3D sensors is disclosed. Each sensor has its own tracker and the filtered data is brought to a common time. One sensor is selected as a master and the tracks of the other sensors are compared with the master to calculate correction factors. This method requires 3D sensors since each sensor has its own complete track. Furthermore, the se
Issing Gregory C.
Saab AB
Young & Thompson
LandOfFree
Bias estimating method for a target tracking system does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Bias estimating method for a target tracking system, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Bias estimating method for a target tracking system will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2846761