Track Maintenance
plots associated
Figure 1: Confirmed track, temporary plot loss, and track termination
Figure 1: Confirmed track, temporary plot loss, and track termination
Figure 1: Confirmed track, temporary plot loss, and track termination
Track Maintenance
Track Maintenance refers to the monitoring and continuation of existing tracks in automatic radar target tracking. After each new radar measurement, the radar tracker must decide whether a track can be updated with a new plot, whether it should temporarily be continued without an associated measurement, or whether tracking must be terminated. The purpose of these procedures is to maintain reliable tracking of targets that are still present while removing invalid tracks from the system in a timely manner.
If no new radar plot can be associated with an existing track during an antenna revolution, this does not necessarily mean that the target has disappeared. For example, the target may have left the radar coverage area or an aircraft may have landed. Alternatively, a plot may have been missed because of unfavorable propagation conditions, obscuration, a weak radar echo, or temporary interference. Errors in associating radar plots with existing tracks may also prevent an actual target measurement from being used to update the track.
Therefore, a single missed detection is usually not sufficient to terminate a confirmed track immediately. Instead, the track may be continued for a limited number of subsequent radar scans, with its position predicted from the most recently estimated motion state. This temporary continuation without a newly associated measurement is called Coasting. It allows target tracking to resume after a brief interruption in measurements as soon as a suitable radar plot is detected again.
During coasting, the target's motion state continues to be predicted using the tracking filter. However, because no new measurement is available for correction, the uncertainty of the estimated target position generally increases. When a Kalman filter is used, this uncertainty is described by the propagated covariance matrix of the state estimate. Consequently, the longer the measurement interruption lasts, the less reliable the prediction becomes. The rate at which uncertainty increases depends particularly on the motion model employed and the assumed uncertainties in the target's motion.
Deciding Whether to Continue a Track
Various criteria can be used to decide whether a track should be continued or terminated. Simple radar trackers use fixed rules based on the number of consecutive missed detections or the number of missed detections within a specified period. More sophisticated systems also consider the estimated track quality, the uncertainty of the position prediction, and the probability of detecting the target under the prevailing conditions.
Common decision methods include:
- Consecutive missed detections: A track is terminated if no suitable plot can be associated with it during a specified number M of consecutive antenna revolutions. A possible setting is, for example, M = 3.
- M-out-of-N rule: The decision is based on a specified number M of missing or available measurements within an observation window of N radar scans. This gives isolated missed detections less influence than prolonged measurement losses.
- Assessment of estimation uncertainty: A track may be terminated if the uncertainty of the state estimate, described by its covariance matrix, exceeds a specified threshold.
- Assessment of track quality: Track continuation may depend on a quality metric that considers, for example, successful measurement associations, missed detections, and the reliability of the motion prediction.
The choice of method and its thresholds depends on the radar's characteristics and the requirements of the tracking task. Terminating tracks too quickly causes them to be lost even during brief measurement interruptions. Conversely, continuing tracks for too long without new measurements may leave outdated or incorrect target information in the system. It also increases the risk that plots received later will be associated with the wrong track. Track maintenance must therefore strike an appropriate balance between maintaining continuity of target tracking and removing invalid tracks promptly.
Track Continuation and Termination
If a suitable radar plot is detected again after one or more missed scans, the track can be updated using that measurement. The tracking filter corrects the predicted motion state and uses the updated estimate for the next position prediction. Target tracking can thus continue without necessarily creating a new track through track initiation.
In a graphical display of the Track History, a missing measured position may be replaced by the target position predicted for that time. This preserves the visible time sequence of the estimated target motion. However, such a position is only a calculated substitute and must not be treated as an actual measured radar plot. In particular, it must not be used as an additional independent radar measurement to correct the tracking filter without properly accounting for its uncertainty.
If a track cannot be reliably updated within the permitted time or number of radar scans, it is terminated. Its designation as a currently tracked target is removed, and the associated track data are deleted or retained for later analysis according to the radar system's requirements. If the same target is detected again at a later time, a new track may have to be established through track initiation. Terminating a track therefore means primarily that automatic target tracking can no longer be continued reliably. It does not provide conclusive evidence that the target has actually left the radar coverage area.