Ultrasonic tracking system

An ultrasonic tracking system is a 3D tracking system that uses ultrasonic sound to measure the position, and often the full position and orientation, of a tracked object such as a head-mounted display, a hand controller, a glove or a stylus. It is also called ultrasonic tracking or acoustic tracking. Emitters send sound at frequencies above the range of human hearing, receivers pick it up, and the system turns the travel time or the phase of the sound into distances between known points, from which it computes position by multilateration. In a 1996 technical report, Tomasz Mazuryk and Michael Gervautz described acoustic trackers as using ultrasonic waves above 20 kHz, with multiple emitters (typically three) and multiple receivers (typically three) of known geometry, because sound can only give the distance between two points. In their 2002 survey of motion tracking, Greg Welch and Eric Foxlin described two families of acoustic ranging: pulsed time-of-flight systems, which time the flight of a brief ultrasonic pulse, and continuous-wave "phase-coherent" systems, which measure the phase shift between the transmitted and received signal.

Ultrasonic tracking needs a line of sight (auditory sight), which is a line of unimpeded travel between source and sensor. It can be built with very small parts: the pen-sized Lincoln Wand of 1966 carried a single receiver, and current designs use MEMS ultrasonic transducers a few millimetres across. Murata has proposed MEMS ultrasonic tracking for styluses and for future smart glasses that would use their built-in microphones as receivers.

Ultrasonic tracking has a long history in virtual reality and augmented reality. Ivan Sutherland's 1968 head-mounted display could use an ultrasonic head position sensor built at MIT Lincoln Laboratory, the Mattel Power Glove of 1989 tracked the hand with ultrasonic transmitters, and the inertial-ultrasonic InterSense IS-900, which its current owner Thales says has thousands of installations, was used in simulators and CAVE displays. Thales announced the end of life of the IS-900 line in May 2025. In 2018 and 2019 ultrasonic controller tracking from Chirp Microsystems appeared on standalone VR headsets, including the HTC Vive Focus Plus.

Reviewed 21 October 2026. Fact-checked the merged page against the Roberts 1966 and Sutherland 1968 papers (full text), Welch and Foxlin 2002, Mazuryk and Gervautz 1996, Pausch 1991, the Deering patent, the archived InterSense IS-600 page and IS-900 release, the current IS-900 page and data sheet, the 1999 Military and Aerospace Electronics report, the TDK, UC Davis, Murata and HTC releases, all cited Road to VR articles, the MilliSonic paper, and the abstracts and Crossref records of the other cited papers. About review dates.

How it works

Time of flight

A pulsed system sends a short burst of ultrasound and measures how long it takes to arrive. Sound travels at roughly one foot per millisecond, so the measured delay converts directly to a range; with ranges from one emitter to several receivers (or from several emitters to one receiver) the system can solve for position by multilateration. Mazuryk and Gervautz named the Logitech 6DOF Ultrasonic Head Tracker and the Mattel Power Glove as examples of time-of-flight trackers. Welch and Foxlin wrote that "all known commercial acoustic ranging systems operate by timing the flight duration of a brief ultrasonic pulse". A practical advantage of pulsed time-of-flight ranging is its handling of echoes: the system can wait for the first pulse to arrive, which must have come by the direct path unless the path is blocked. Because sound is slow, there is a large time gap between the direct pulse and the first reflection, so the method works for sound even though it does not work for radio or light. Roberts used the same reasoning in 1966: in the Lincoln Wand, a received pulse stopped a counter started when the transmitter fired, and "if any reflections are seen by the receiver they occur after the straight path reception and are therefore ignored".

The transducers need to be close to omnidirectional so that the signal can be detected whatever the orientation of the tracked object, which means small speakers and microphones with active surfaces a few millimetres across. Small transducers are less efficient, so most systems drive highly resonant transducers at their resonant frequency. The received waveform then "rings up" over about 10 cycles; a typical envelope-peak detector triggers about 10 cycles (about 90 mm) late. Welch and Foxlin note that detecting on the second or third cycle instead greatly reduces the risk of multipath reflection, and they describe this as the crux of how InterSense's ultrasonic ranging stays accurate at longer ranges than others. (Foxlin was the founder and chief technology officer of InterSense.)

Phase-coherent tracking

A continuous-wave system transmits a steady tone and measures the phase of the received signal. The phase changes continuously as the path length changes, so the measurement has no inherent delay, but it only reveals the position within one wavelength. To obtain absolute distance the system must know the starting distance and count the whole cycles gained or lost as the object moves. A phase difference of 360 degrees is equivalent to a distance of one wavelength; Mazuryk and Gervautz wrote that because the method delivers relative data whose error tends to accumulate with time, "development of PC trackers was relinquished". Sutherland's 1968 sensor worked this way, and his paper describes the resulting ambiguity: at 40 kHz the wavelength of sound in air is about one third of an inch, so each measurement was ambiguous at 1/3 inch intervals, and any wrong guess of the starting path length showed up as a constant "initialization error". Welch and Foxlin suggested that multipath reflections, which make the amplitude and phase of a continuous-wave signal in a room vary "drastically and unpredictably" as the receiver moves, might be why no successful phase-coherent tracker had been developed. The method has its own article, Phase coherent ultrasonic tracking.

Position and orientation

One range measurement places a point on a sphere around an emitter or receiver; three or more ranges to known points fix the point in space. Putting several emitters or receivers on one rigid object gives the positions of several points on it, from which the orientation can be derived. Sutherland's sensor, with three transmitters on the headset and four receivers on the ceiling, made twelve measurements from which, in Sutherland's words, "head-position information can be derived"; the paper defines the head position sensor's job as measuring both the position and the orientation of the user's head. The Power Glove, with two transmitters on the back of the hand, reported position plus a roll angle in one of twelve steps.

Redundant ranges can also catch errors. The Lincoln Wand used four transmitters although three distances are enough to fix a point, because the fourth gave a geometric check: Roberts wrote that a bad distance reading is usually either very small or very large, and that a reflection received when the direct path is blocked "is very consistent when it occurs and can only be detected by a geometric check". Placing the four transmitters at the corners of a square made both the position and the error check simple to compute.

Finding the rotation that best maps a set of known points on an object onto a set of measured points is the problem solved by the method Wolfgang Kabsch published in 1976, which determines "a best rotation of a given vector set into a second vector set by minimizing the weighted sum of squared deviations". Sutherland's report on the ultrasonic positioning system used with his 3D display did not mention any use of the Kabsch algorithm, which was published eight years later.

Where the receivers sit determines the arrangement. In an outside-in layout the microphones are fixed in the room and the emitters ride on the tracked object, as in Sutherland's display and the Power Glove. The Lincoln Wand reversed the roles of sender and listener, with four fixed transmitters and the receiver in the hand-held wand. The Constellation system described by Foxlin, Harrington and Pfeifer in 1998 used "inside-out ultrasonic tracking" in which ranges are measured between the tracked unit and a constellation of wireless beacons, which allowed tetherless tracking over a building-wide range. In the consumer controller trackers of 2018 and 2019 the receivers sat on the headset itself, an inside-out arrangement that tracked the controllers relative to the head.

Limitations

Ultrasonic ranging depends on the speed of sound, which varies with temperature, humidity and air currents, and outdoors wind also affects accuracy. Welch and Foxlin give a rule of thumb that the speed of sound changes by about 0.1 percent per degree Fahrenheit, about a one-millimetre error per degree Fahrenheit at a range of one metre. A Sun Microsystems patent by Michael Deering, for example, states that its triangulation of tracked glasses and a 3D mouse "takes into account the speed of sound given the air temperature and other factors".

Frequency is a trade-off. Most ambient noise falls off at higher frequencies, and shorter wavelengths give higher resolution, but attenuation of sound in air "starts to play a significant role by 40 kHz and becomes the dominant factor in limiting range by 80 kHz, depending on humidity". The Lincoln Wand's receiver amplifier was tuned to 50 kHz, which Roberts wrote "rejects most room noises".

Reverberation limits the update rate. Depending on the room and the tracking volume, a system may have to wait between 5 and 100 ms for echoes of one measurement to die out before starting the next, giving update rates as low as 10 Hz. The time for the sound to reach the receivers adds latency of about one millisecond per foot of range, usually well under 15 ms. The Lincoln Wand illustrates the arithmetic: allowing time for 11-foot reflections to die out, it pulsed its four transmitters at 10 ms intervals, so a full cycle took about 40 ms and the position was sampled 25 times a second. The Logitech 6DOF tracker that Mazuryk and Gervautz used as a typical example had a range of 1.5 m with a 100-degree cone of coverage, an update rate of 50 Hz, a lag of 30 ms and an accuracy of 2 percent of the distance from the source and 0.1 degree in orientation.

Acoustic trackers need a line of sight between emitters and receivers, although they are somewhat more tolerant of occlusion than optical trackers because sound can travel around obstacles. They are also sensitive to other ultrasonic noise: Welch and Foxlin wrote that they had "yet to see a purely acoustic tracker that doesn't go berserk when you jingle your keys". Sutherland chose continuous-wave ultrasound partly to avoid confusion from pulsed noise "such as typewriters produce", which had caused difficulty for the pulsed Lincoln Wand. Mazuryk and Gervautz listed the line-of-sight restriction, acoustic interference from noise or echoes, and low update rates as the disadvantages of time-of-flight trackers, against their light weight, small size, relatively low price (from US$1,000 at the time) and freedom from magnetic interference. Consumer controller trackers show the line-of-sight limit in practice: in Road to VR's hands-on with HTC's ultrasonic controller kit, controllers held behind the user's back were likely to be lost by the system.

Welch and Foxlin concluded that most of these weaknesses can be addressed by a hybrid system that combines acoustic sensors with sensors of complementary character, such as inertial sensors. The InterSense IS-600 Mark2, IS-900 and Chirp's controller tracking for standalone headsets all fuse ultrasonic ranges with data from an inertial measurement unit.

History

Lincoln Wand

The Lincoln Wand was an ultrasonic position sensor at MIT Lincoln Laboratory, connected to the TX-2 computer; Lawrence G. Roberts wrote that it had been operational since April 1966. Sutherland's 1968 paper cites a June 1966 Lincoln Laboratory report by L. G. Roberts on the device and describes it as a pulsed ultrasonic system. Roberts's paper at the AFIPS Fall Joint Computer Conference in November 1966 describes an ultrasonic position-sensing device that let the computer read the x, y and z coordinates of the tip of a pen-sized wand. Four transmitters at the corners of a square, which could be mounted flush on the front panel of a display, each sent a 20-microsecond burst of energy band-limited between 20 and 100 kHz; the receiver in the wand stopped a counter when the pulse arrived. A 1 MHz counting rate resolved distance to 0.014 inch, and a 13-bit counter allowed measurements of up to 9 feet.

Lincoln Wand specifications as given by Roberts (1966)

Specification	Value
Working space	4 x 4 x 6 feet
Resolution	0.02 inch (12 bits per axis)
Sampling rate	25 per second
Computer time	1 percent
Absolute accuracy	0.2 inch (not measured; Roberts's estimate)

Roberts proposed using the wand's third coordinate to draw lines and curves in three dimensions, to move objects "stuck" to the wand, and to point at menus and parts libraries placed around the display, beyond the edge of the screen. He gave US$1,500 as the cost of the ultrasonic equipment and wrote that expansion to four or five wands sharing the same pulsing logic was being planned.

Sutherland's head-mounted display

Sutherland's display (Ivan Sutherland's head-mounted 3D display) could use either of two head position sensors, one mechanical and one ultrasonic. The mechanical arm hanging from the ceiling was "rather heavy and uncomfortable to use", and Sutherland wrote that it had been built "to have a sure method of measuring head position". The continuous-wave ultrasonic sensor was designed and built at the MIT Lincoln Laboratory by Charles Seitz and Stylianos Pezaris. Three transmitters, sending ultrasound at 37, 38.6 and 40.2 kHz, were attached to the head-mounted optical system, and four receivers were mounted in a square array in the ceiling. Each receiver fed an amplifier and three filters, so the phase changes along twelve sound paths could be measured, each read by the computer as a separate five-bit number; the computer counted major changes in phase to follow motions of more than one wavelength. Sutherland chose continuous waves because inexpensive narrow-band transducers were available and to avoid the pulsed-noise problems of the Lincoln Wand.

In preliminary experiments at Lincoln Laboratory in late 1966 and early 1967, Sutherland wrote, the ultrasonic sensor "operated well enough to measure head position for a few minutes before cumulative errors were objectionable". For the later system he reported "encouraging results" from programs written to resolve the measurement ambiguities, but wrote that "a full report on the ultrasonic head position sensor is not yet possible". The same paper states that the initial equipment allowed a working volume of head motion about six feet in diameter and three feet high, within which the user could turn completely around and tilt the head up or down about forty degrees. It sets a target resolution of 1/100 inch and one part in 10,000 of rotation, and says the sensor should have an accuracy on the order of one tenth of an inch.

Power Glove and early low-cost VR

The Power Glove, manufactured by Mattel under licence from Abrams-Gentile Entertainment and sold as a Nintendo Entertainment System peripheral, used ultrasonics for hand position and variable-resistance material for finger bend. Randy Pausch noted that the earlier VPL Research DataGlove used a Polhemus magnetic tracker for hand position, and that neither of the DataGlove's technologies could easily be mass-produced. The glove had two ultrasonic transmitters on the back of the hand and three wall-mounted receivers arranged in an L-shape. It communicated within ten to fifteen feet of the receivers when pointed toward them; Mattel said it was usable up to about 45 degrees away from the receivers, and within five to six feet its x, y, z position was accurate to within 0.25 inches. Roll was reported in one of twelve positions. Pausch used the glove as the hand input of a complete VR system built for under US$5,000 and presented at CHI in 1991; he gave the glove's retail price as 70 to 100 dollars and wrote that Mattel reported more than 40 million dollars of Power Glove revenue in 1989.

Dedicated commercial time-of-flight trackers were also sold. Mazuryk and Gervautz's 1996 report illustrates the Logitech 6DOF Ultrasonic Tracker and quotes its specification (above) as typical of acoustic time-of-flight trackers. A 1992 survey of position trackers for VR by Kenneth Meyer, Hugh Applewhite and Frank Biocca in Presence discussed acoustic implementations alongside mechanical, optical and magnetic ones. Ultrasonic head tracking was also used for head-tracked stereo on desktop displays. A Sun Microsystems patent by Michael Deering, granted in 1995, describes stereo shuttered glasses with embedded ultrasonic microphones and a 3D mouse with three microphones, both tracked by a head tracker with three ultrasonic speakers mounted on the stereo CRT; the travel time of the sound was used to triangulate the 3D positions of the glasses and the mouse.

Constellation, IS-600 and IS-900

By 1999 InterSense sold the IS-600 Mark2, which obtained orientation from a miniature solid-state inertial measurement unit and used ultrasonic time-of-flight distance measurements to keep track of changes in position. InterSense described the method as "accelerometry with ultrasonic drift correction, not just the pure time-of-flight trilateration used by others", and claimed update rates of up to 150 Hz and angular motion prediction of up to 50 ms. In 1999 Military and Aerospace Electronics reported that Lockheed Martin had selected the off-the-shelf IS-600, which combined inertial and wireless ultrasonic tracking, for a proof-of-concept indoor combat training system for the US Marine Corps, with sensors on the ceilings and small transmitters on the trainees' helmets.

In 1998 Eric Foxlin, Michael Harrington and George Pfeifer presented Constellation, a tracking system for augmented reality and virtual set applications "based on an inertial navigation system aided by ultrasonic time-of-flight range measurements to a constellation of wireless transponder beacons". An extended Kalman filter working on individual range measurements let the inertial sensors reject corrupt ranges and perform smoothing and prediction, while the screened ranges corrected the drift of the inertial system. The authors reported that the implemented system's accuracy matched their simulation, and that the simulation predicted accuracies better than 2 mm with further compensation for transducer misalignment.

Constellation was brought to market as the InterSense IS-900. It uses a ceiling-mounted array of reference emitters in strips that clip to ceiling grids, and it keeps latency low by fusing its relatively slow acoustic range measurements with inertial sensors. InterSense announced the IS-900 series on 9 August 1999 in four versions: the VET (Virtual Environment Tracker) for CAVEs and cubes, the VWT (Virtual Workbench Tracker), the LAT (Large Area Tracker) for room and wall displays and the SCT (Studio Camera Tracker) for production studios. A related acoustic-inertial product, the InterSense SimTracker, was used in the AVCATT-A military simulator, where it offered a prediction option to address rendering latency. InterSense later became part of Thales, and the 2024 Thales data sheet specifies the IS-900 as follows:

Specification	IS-900 (manufacturer data)
Tracking volume	Up to 20 m2 with 12 SoniStrips (SimTracker); up to 140 m2 with 84 SoniStrips (VETracker)
Resolution	0.75 mm; 0.05 degrees
Static accuracy	2.0-3.0 mm; 0.25 degrees RMS in pitch and roll, 0.50 degrees RMS in yaw (measured in a controlled environment)
Update rate	180 Hz nominal
Latency	4 ms typical
Prediction	Up to 50 ms
Maximum number of devices	4 (SimTracker); 7 (VETracker)

Thales describes the IS-900 as used in CAVEs, Powerwalls, flight and weapons simulators and film and video production, with "1,000's of installations worldwide". It announced on 1 May 2025 that the IS-900 product line was going end of life, with a last-time-buy option through 31 December 2025.

Spread-spectrum audio

In 2002 Welch and Foxlin reported that Vallidis and Bishop at the University of North Carolina at Chapel Hill were working on body-relative limb tracking using spread-spectrum audio: small speakers transmit pseudorandom signals and signal processing finds the delay to small microphones. The approach resists environmental noise, spreads the energy so the audible sound is reduced "to a whisper", and because low-frequency audio passes around the body it is relatively immune to occlusion.

Controller tracking for standalone headsets

In 2018 ultrasonic tracking returned to consumer VR hardware as a way to give the controllers of standalone headsets six degrees of freedom. On 26 December 2017 Road to VR reported that Pico's new Pico Neo standalone headset would track its 6DoF controllers with ultrasonic technology, which Pico said needed no external cameras, trackers or wires. At CES 2018 Pico showed the headset with an ultrasonic controller solution from Chirp Microsystems; the headset had three ultrasonic receivers, one on each side and one at the top centre. Road to VR found latency "a pretty significant issue" in that pre-consumer build. In April 2018 Road to VR reported that Qualcomm's Snapdragon 845 VR Development Kit would offer ultrasonic 6DOF controller tracking, with six ultrasonic emitters or receivers on the headset, three on each side; it noted that ultrasonic tracking had historically been "considered insufficient as a head-tracking solution for high-end VR headsets".

In October 2018 HTC offered developers a 6DOF controller kit for the HTC Vive Focus, consisting of two controllers and a large module mounted on the headset. It combined ultrasonic tracking with IMUs over a field of 180 degrees horizontally and 140 degrees vertically, with "high accuracy" up to one metre from the headset. In Road to VR's hands-on, accuracy "seemed good", but precision was "noticeably not at the 'sub-mm' bar set by high-end tracking systems", with visible jitter when the controller was held still, and positional latency was higher than on leading controller tracking systems. A February 2019 Road to VR test of a later build, with a faceplate receiver connected to the headset by USB, found jitter "much less" than on the Pico Neo a year earlier, though still more than on Windows VR headsets; Chirp claimed a tracking volume of more than 180 degrees extending to a maximum of 1.2 m.

On 8 January 2019 TDK, whose group by then included Chirp Microsystems, announced the technology as Chirp SonicTrack, an inside-out system that "fuses ultrasonic and inertial sensor data" to track a handheld controller in six degrees of freedom, built on Chirp's CH-101 ultrasonic transceiver. TDK claimed sub-millimetre precision, a field of view of more than 240 degrees and a power draw of a few milliwatts, which the body of the release puts at about a hundred times less than an optical tracking system (its subtitle gives a different ratio), and named the Vive Focus developer kit as the first VR system to use it. On the same day Chirp and Qualcomm Technologies announced a collaboration to support SonicTrack on the Snapdragon Mobile XR platform. HTC announced the Vive Focus Plus on 21 February 2019 with "two ultrasonic 6DoF controllers", which Road to VR identified as created by Chirp Microsystems; in March 2019 HTC set availability for 15 April 2019 at US$799 for enterprise customers, initially in 25 markets.

MEMS ultrasonic transducers

Chirp Microsystems grew out of research at UC Davis and UC Berkeley. According to UC Davis, which reported TDK's acquisition of the Berkeley-based start-up in March 2018, its micro-electromechanical sensors send out ultrasonic pulses and time the returning echo, sense ranges from several centimetres to several metres, and use very little power because the transducers are only a few millimetres across; it listed augmented and virtual reality systems among the applications. UC Davis professor David Horsley was a co-founder and chief technology officer, and the technology was developed with Bernhard Boser of UC Berkeley.

The underlying devices are piezoelectric micromachined ultrasonic transducers (PMUTs). In a 2011 IEEE Sensors Journal paper, Richard Przybyla, Horsley, Boser and colleagues described an in-air rangefinder built around an aluminium nitride (AlN) membrane 400 micrometres in diameter, fabricated with a low-temperature process compatible with processed CMOS wafers. Using pulse-echo time of flight with a single transducer for transmit and receive, it worked from 30 to 450 mm at up to 375 samples per second, with random noise of 1.3 mm at maximum range. A 2015 paper in the IEEE Journal of Solid-State Circuits extended this to a 3D rangefinder on a chip: an array of AlN MEMS transducers with a custom 0.18 micrometre CMOS readout chip localized targets over a field of view of plus or minus 45 degrees up to 1 m away, with root-mean-square position errors at 0.5 m of 0.4 mm in range and 0.2 and 0.8 degrees in the two angles, and a power dissipation of 400 microwatts at 30 frames per second for a 1 m maximum range. TDK describes Chirp's transducers as "piezoelectric MEMS ultrasonic transducers".

Murata is also developing a piezoelectric MEMS ultrasonic transducer (pMUT) for tracking. In a 2024 article it described a time-of-flight approach in which ultrasound from the pMUT is received by MEMS microphones, and multiple transducers' one-dimensional distances are combined to calculate an object's position. The article showed the Maliang Magic Pencil, a stylus from Sensoryx that combines ultrasound, optical and inertial tracking, and named watches, pens, rings and tools such as training tools, surgical scalpels and stylus pens as further uses. It presented smart glasses as a future application that could use their multiple microphones for position tracking "even under intense light condition", and stated that ultrasound performs consistently in all lighting conditions, including direct sunlight. Murata gave no numerical accuracy or latency figures.

Acoustic tracking with consumer devices

Research groups have also tracked phones and headsets with the speakers and microphones they already contain, often at frequencies near the top of the audible range rather than true ultrasound. The CAT system presented at MobiCom 2016 by Wenguang Mao and colleagues at the University of Texas at Austin was designed to let a user play games, interact with VR and AR headsets and control smart appliances by moving a smartphone in the air. Several speakers transmit inaudible sounds at different frequencies, and a distributed frequency-modulated continuous wave (FMCW) estimates the absolute distance between an unsynchronized transmitter and receiver; the system combines these estimates with Doppler shifts and inertial measurements. MilliSonic, presented at CHI 2019 by Anran Wang and Shyamnath Gollakota of the University of Washington, emits 45 ms chirps from 17.5 to 23.5 kHz through a smartphone speaker to a four-microphone array; the authors note that smartphones have "a limited inaudible bandwidth of 7 kHz between 17-24 kHz". They reported a median accuracy of 0.7 mm in 1D and 2.6 mm in 3D, and demonstrated 3D tracking of VR headsets using a smartphone as a beacon and of a Google Cardboard viewer with a small microphone array.

Ultrasonic ranging is also used for indoor positioning outside VR. The Cricket location-support system presented at MobiCom 2000 by Nissanka Priyantha, Anit Chakraborty and Hari Balakrishnan used concurrent radio and ultrasonic signals from beacons to let mobile devices infer their distance and learn their location inside a building, with each device built from off-the-shelf parts for under US$10. Mike Hazas and Andy Hopper argued in 2006 that ultrasonic location systems were limited by their narrowband transducers, and built broadband ultrasonic transmitters and receivers for two positioning systems. The TELIAMADE system, described in 2013 by C. Medina, J. C. Segura and A. de la Torre of the University of Granada, is another indoor positioning system rather than a VR tracker. Its wireless nodes carry a radio module and an ultrasound module and share the ultrasonic channel by time-division multiplexing; the authors estimate time of flight on a 40 kHz carrier with a quadrature detector, refine it with parabolic interpolation of the matched-filter envelope and with the signal phase, and report pseudorange errors below 2 mm RMSE at distances of 2 to 6 m and an average location RMSE of 9.6 mm.

Examples

System	Year	Arrangement	Ranging method	Reported performance
Lincoln Wand	1966	Four transmitters at the corners of a square, one receiver in a pen-sized wand	Pulsed time of flight	0.02 inch resolution, 25 samples per second, 4 x 4 x 6 foot working space
Ultrasonic head position sensor for Ivan Sutherland's head-mounted display	1968	Three transmitters (37, 38.6 and 40.2 kHz) on the headset, four receivers in the ceiling	Continuous-wave phase	Full report "not yet possible" at publication
Power Glove	1989	Two transmitters on the glove, three wall-mounted receivers in an L-shape	Ultrasonic ranging	0.25 inch within five to six feet of the receivers
Logitech 6DOF Ultrasonic Head Tracker	By 1996	Emitters and receivers of known geometry	Pulsed time of flight	1.5 m range, 100-degree cone, 50 Hz, 30 ms lag, 2 percent of distance and 0.1 degree accuracy
Constellation / InterSense IS-900	1998 (research system), 1999 (product)	Tracked unit and ceiling-mounted ultrasonic beacons, fused with inertial sensors	Time of flight with an extended Kalman filter	IS-900 data sheet: 2.0-3.0 mm static accuracy, 180 Hz nominal, 4 ms typical latency
InterSense IS-600 Mark2	By 1999	Inertial measurement unit with ultrasonic drift correction	Time of flight	Up to 150 Hz; angular prediction up to 50 ms
Chirp SonicTrack (HTC Vive Focus controller kit, HTC Vive Focus Plus)	2018-2019	Ultrasonic transducers in the controllers and on the headset, fused with IMU data	Ultrasonic ranging	HTC kit: 180 by 140 degrees, "high accuracy" up to one metre; TDK claims sub-millimetre precision
TELIAMADE	2013	Wireless nodes with a radio module and an ultrasound module, sharing the channel by time-division multiplexing	Time of flight on a 40 kHz carrier, located by multilateration	Average location RMSE of 9.6 mm
MilliSonic	2019	Smartphone speaker as the source, four-microphone array as the receiver	FMCW chirps at 17.5-23.5 kHz with phase tracking	Median 0.7 mm (1D) and 2.6 mm (3D)

See also

• Phase coherent ultrasonic tracking
• Lincoln Wand
• Ivan Sutherland's head-mounted 3D display
• Sword of Damocles
• InterSense
• Multilateration
• Positional tracking
• Inside-out tracking
• Outside-in tracking
• Electromagnetic tracking
• Optical tracking
• Sensor fusion
• 3D stylus
• Comparison of tracking systems