Track_while_scan - Wikiparency

# Track while scan **Track-while-scan** (TWS) is a radar operating mode that keeps searching a volume of space for new targets while simultaneously maintaining an updated position and velocity estimate, a track, on every target already detected, rather than giving the whole sensor over to one job at a time. It sits as a compromise between two older extremes: a pure search radar, which reports a stream of unconnected detections with no memory of where a target was a moment before, and a pure single-target tracker, which points a narrow beam continuously at one target and leaves the rest of the sky unwatched while it does. A three.js companion sketch elsewhere on the site renders the recursive filter this mode relies on, showing how the uncertainty around a predicted target position widens the longer the beam goes without revisiting it and narrows again the instant a fresh detection arrives. Because the same scan that searches also updates every open track, track-while-scan output feeds directly onto an operator's [[Radar_display]], typically as a set of smoothed, labelled tracks laid over the raw plot picture, while the tracking machinery itself draws on the general estimation methods described under [[Radar_tracker]]. The mode became practical once radars carried enough onboard computing to hold several targets' state estimates in memory and update the right one after every beam position, and it remains the ordinary way a modern search-and-track radar, mechanically scanned or electronically steered, reports on the objects it is watching. ## Background Before a radar could hold a running estimate on more than one target at a time, an operator had two separate tools and had to choose between them. A search radar swept a wide volume and reported each detection as an isolated plot, a position and perhaps a Doppler-derived closing rate, with nothing connecting one scan's plot to the next; deciding by eye which blips from successive sweeps belonged to the same aircraft was left to a human plotter or a separate piece of equipment entirely. A dedicated tracking radar went the other way, narrowing its whole beam onto a single target and holding it there with a continuous, closed loop, of the kind a [[Monopulse_radar|monopulse]] fire-control set uses, at the cost of going effectively blind to everything else in the sky while it did so. Track-while-scan grew out of the demand, sharpest in Cold War-era air-defence and interceptor radars built to find and follow more than one aircraft at once in [[Ultra_high_frequency|UHF]]- and microwave-band search volumes, for a single sensor that could do both jobs from the same beam.[^tws-origin] Interceptors sent up against high-altitude bombers crossing the [[Stratosphere|stratosphere]] for the [[United_States_Air_Force|United States Air Force]] were among the first users pressing for the capability, since a crew holding a single-target lock on one bomber had no way to see a second one closing on them at the same time.[^usaf-interceptor] What made the mode practical was not a new antenna or a new waveform but enough onboard computing, at first purpose-built analogue and digital hardware and later a general-purpose processor, to hold a separate state estimate, an expected position and velocity, for every target the radar had accepted as real, and to update the correct one each time the beam swept past it. Before that capacity existed, dividing a radar's revisits between search and several independent tracks was possible in principle but impractical with the equipment of the day.[^tws-compute] ## Track while scan In practice, the radar's normal search scan supplies the raw material: whichever pattern the antenna sweeps, mechanically rotating or, in a modern [[Phased_array|phased array]] or [[Active_electronically_scanned_array|active electronically scanned array]], stepped electronically from one direction to the next, occasionally happens to illuminate a target already carrying an open track. A tracking function tests each new detection against every open track's predicted position; a detection that falls inside a track's gate, the region around the predicted position sized to the estimate's own uncertainty, updates that track through a recursive filter, folding the new measurement together with the prediction to produce a new position, velocity and a smaller uncertainty than either alone. A detection that falls outside every existing gate either starts a new, tentative track or is discarded as clutter, depending on how the system is tuned to trade missed targets against false ones. Between illuminations, with no new measurement to work from, each track's predicted position simply advances along its last known velocity, and the gate around that prediction grows, since more time has passed for the target to have manoeuvred away from where a straight-line prediction would put it; the next scan that reaches the target's true position shrinks the gate back down as soon as a fresh detection lands inside it. This predict-then-update cycle, applied independently to every track a radar is holding, is the same idea a Kalman filter formalises for a single one, and it is what lets one radar hold a dozen or more tracks current using only the fraction of a second its beam spends on each per revisit, rather than the whole of its time on one target the way a continuous tracker would need. Not every detection a track-while-scan radar accepts is real. Ground and sea clutter, and the false alarms that random receiver noise occasionally produces by crossing a detection threshold, both compete with genuine targets for a track; a radar tuned to a [[Constant_false_alarm_rate|constant false-alarm rate]] adapts that threshold to the local clutter level so a busy patch of sea does not spawn a wave of spurious tentative tracks, and [[Moving_target_indication|moving-target indication]] filtering upstream removes much of the near-zero-velocity [[Clutter_(radar)|clutter]] before the tracker ever sees it. Even so, a track-while-scan system routinely carries several logical states for each candidate at once: a tentative track that has not yet earned enough confirmed hits to be shown to an operator, a firm track presented on the display, and a coasting track carried through one or more misses on the strength of its prediction alone before it is finally dropped. That bookkeeping is a separate problem from, but built directly on top of, the predict-and-update cycle described above. Because every open track only has to be revisited, not watched continuously, the number of targets a track-while-scan radar can carry is set chiefly by how many gates its processor can maintain and how quickly the antenna returns to each one, not by how many receivers or beams the hardware happens to have; a mechanically scanned antenna that turns once every few seconds can still support many simultaneous tracks provided the tracker's gates stay wide enough to bridge the interval.[^track-capacity] An electronically scanned array sharpens this further, since it can interleave brief dwells on several tracks with its ordinary search frame in almost any order the scheduling software chooses, rather than waiting for a slow mechanical rotation to bring the beam back around; upgrading an older search radar to an active electronically scanned array is often marketed chiefly as a track-while-scan improvement rather than a search-range one.[^aesa-upgrade] The trade the mode makes is precision for coverage. A dedicated single-target tracker, watching one target with every available beam and often deriving angle from a monopulse comparison of a single pulse rather than from successive scans, resolves that one target's position and velocity more tightly than any track-while-scan estimate revisited only occasionally can. Track-while-scan accepts a coarser, more slowly updated picture of any one target in exchange for a simultaneous, continuously refreshed picture of every target the search beam finds, which is the picture an air-traffic controller, an airborne intercept crew, or a ship's air-defence operator actually needs, typically presented as a set of smoothed track symbols and velocity vectors overlaid on a [[Plan_position_indicator|plan-position]] style display rather than as raw, scan-to-scan blips. ## Microsims This article carries no p5.js sketch of its own. A three.js companion instead renders the predict-and-update cycle described above directly: a single track's gate is shown widening scan by scan while no detection lands inside it, then snapping back down the moment one does, the same geometry a track-while-scan filter works through on every open track at once. *Try:* in the [[Radar]] sketch, drag the target's range control and watch how many rotations of the sweeping beam pass before it revisits that bearing; a track-while-scan filter must predict through exactly that gap between illuminations, since it is the antenna's rotation rate, not the pulse-repetition frequency, that sets how often any one track can be updated. *Try:* in the [[Doppler_effect]] sketch, raise the source speed and watch the two observer markers' frequency readouts diverge; the same shift, read off a target's echo, gives a track-while-scan filter the radial-velocity term that lets its prediction follow a target's real heading rather than only the position where it was last seen. ## Wikipedia : Wikitube **Strict pair:** [Wikipedia](https://en.wikipedia.org/wiki/Track_while_scan) : [Wikitube](https://en.wikitube.io/wiki/Track_while_scan) Skeleton mirrored at revision 1373342868. Prose, emphasis and the microsims are Wikitube's own. ## See also - [[Radar_tracker]] - [[Radar_display]] - [[Monopulse_radar]] - [[Phased_array]] - [[Active_electronically_scanned_array]] - [[Plan_position_indicator]] - [[Moving_target_indication]] ## References The predict-and-update cycle described above is the same recursive-estimation idea used throughout tracking and control, and, like the general Kalman-filter formalism it draws on, is standard estimation theory and is not separately footnoted here, per the Wikitube style guide's §6.1. The sub-manual section this topic plans to, manual 07 §6.2, treats the same recursive least-squares idea but cites only a source outside this article's approved book list; no claim below is drawn from it. [^tws-origin]: Citation needed: the paper, program record or patent that first names "track-while-scan" as a radar mode and dates its first operational use would confirm the origin claimed here. [^usaf-interceptor]: Citation needed: a named United States Air Force interceptor fire-control radar and its in-service date would support the claim that stratospheric bomber defence drove early demand for the mode. [^tws-compute]: Citation needed: a technical history of an early track-while-scan radar's tracking hardware or computer would confirm when onboard processing first became capable of holding several simultaneous tracks. [^track-capacity]: Citation needed: a named modern radar's documented maximum simultaneous track-while-scan track count would support a concrete figure in place of the general claim made here. [^aesa-upgrade]: Citation needed: a manufacturer's or operator's account of a specific active-electronically-scanned-array radar upgrade marketed on its track-while-scan improvement would confirm this claim. **Further reading** — the open textbooks this article draws on, since the pair carries no separate Further reading heading of its own: - Christian Tiberius; Max Mulder. *Engineering Signal Analysis: From Fourier to filtering: Theory*. 2026. Open Textbook Library: https://open.umn.edu/opentextbooks/textbooks/engineering-signal-analysis-from-fourier-to-filtering-theory . CC BY. - Don Johnson. *Fundamentals of Electrical Engineering I*. 2014. Open Textbook Library: https://open.umn.edu/opentextbooks/textbooks/fundamentals-of-electrical-engineering-1 . CC BY. - Michael Stiber; Bilin Stiber; Eric Larson. *Signal Computing: Digital Signals in the Software Domain*. 2020. Open Textbook Library: https://open.umn.edu/opentextbooks/textbooks/signal-computing-digital-signals-in-the-software-domain . CC BY-SA. - Allen Downey. *Think DSP: Digital Signal Processing in Python*. 2012. Open Textbook Library: https://open.umn.edu/opentextbooks/textbooks/think-dsp-digital-signal-processing-in-python . CC BY-NC. - Steven Ellingson. *Radio Systems Engineering, Revised First Edition*. 2023. Open Textbook Library: https://open.umn.edu/opentextbooks/textbooks/radio-systems-engineering . CC BY-NC. - John Dyer; Chad Davis. *Measurement and Instrumentation: An Introduction to Concepts and Methods, 1st Edition*. 2020. Open Textbook Library: https://open.umn.edu/opentextbooks/textbooks/measurement-and-instrumentation-an-introduction-to-concepts-and-methods . CC BY-NC-SA.