Toggle search
Toggle menu
5.5K
6.1K
231
53.7K
VR & AR Wiki
Toggle preferences menu
Toggle personal menu
Not logged in
Your IP address will be publicly visible if you make any edits.

Inside-out tracking

vrarwiki.com · VR & AR Wiki
More actions
(Redirected from Inside-Out Tracking)
See also: Terms and Technical Terms
See also: Markerless inside-out tracking, Positional tracking and Outside-in tracking
The front of a Meta Quest 3, showing the outward-facing cameras used for inside-out tracking
Reviewed 17 October 2026. Checked every cited source (Welch & Foxlin 2002, UNC 1990 papers, HiBall, MonoSLAM/PTAM/MSCKF/MEgATrack/Holzwarth/Abdlkarim abstracts and DOIs, Meta Insight and Melim posts, Microsoft, Apple, Samsung, Valve, HTC, PICO, Magic Leap, Sony and press articles) against the article's claims and corrected unsupported wording. About review dates.

Introduction

Figure 1. A Windows Mixed Reality headset, an example of an HMD using markerless inside-out tracking via cameras visible on the front (image: www.wareable.com)

Inside-out tracking (also called inside-looking-out tracking) is a method of positional tracking used in virtual reality (VR), augmented reality (AR), and mixed reality (MR) systems. It lets devices such as head-mounted displays (HMDs) and motion controllers determine their position and orientation in 3D space using sensors located on the device itself, looking outward at the surrounding environment.[1]

It contrasts with outside-in tracking, where external sensors (for example cameras or laser emitters) are placed in the environment to track sensors or markers located on the HMD or controllers. In inside-out tracking, the cameras or other sensors are on the moving object and the reference points are fixed in the surroundings, making the system inherently egocentric.[1][2] The term was in use in VR research by 1990, when a University of North Carolina team described a head tracker that "adopts an 'inside-out' tracking paradigm", with cameras on the user's helmet viewing infrared LEDs on the ceiling.[3]

Markerless, camera-based inside-out tracking is used by standalone headsets such as the Meta Quest line, by tethered headsets such as PlayStation VR2, and by mixed reality headsets including the Apple Vision Pro and the Samsung Galaxy XR, and by Valve's Steam Frame.[4][5][6][7][8] The approach removed the external tracking hardware that earlier consumer systems required: the Oculus Rift CV1 used external Constellation sensors, and the original HTC Vive used Lighthouse base stations. That made room-scale VR simpler to set up.[4][9][10]

Terminology

Greg Welch and Eric Foxlin, in their 2002 survey of motion tracking in IEEE Computer Graphics and Applications, describe the choice for optical systems as whether to put the light sources on the moving target and the sensors in the environment ("outside-looking-in") or the reverse ("inside-looking-out").[1] They caution that the labels can be misleading. Their example is the Vulcan measurement system from Arc Second, which mounted optical sensors on the target and used spinning light sources in the room that swept planes of light across them; although the sensors "look outward", the system has the orientation sensitivity of an outside-looking-in design. In their view, the actual distinguishing factor is whether bearing angles to reference points are measured from the outside or from the inside.[1]

The same ambiguity appears in consumer VR. Valve's Lighthouse system places photosensors on the headset and controllers and sweeps the room from fixed base stations, and PCWorld's Hayden Dingman noted in 2018 that the HTC Vive "technically" also uses inside-out tracking, while the inside-out tracking of Windows Mixed Reality headsets is done with cameras and no base stations.[10] Manufacturers use the term for the second kind, with cameras and inertial sensors on the device and no external hardware. Meta described Oculus Insight as an "inside-out" system that replaced external sensors attached to a PC, and Google's documentation for Daydream standalone headsets equated its WorldSense 6DoF tracking with "inside-out tracking".[4][11]

How it works

An inside-out tracking system observes the external world from the perspective of the device being tracked. As the device moves, the sensors detect changes relative to the environment, allowing the system to calculate the device's updated position and orientation (six degrees of freedom: three translational, three rotational). This calculation happens in real time so that the virtual environment can respond as the user moves.

There are two main approaches.

Marker-based inside-out tracking

Figure 2. Valve's "Room" prototype, shown at Steam Dev Days in January 2014: cameras on the headset tracked 2-dimensional barcode fiducial markers covering the walls and ceiling.

This approach relies on placing artificial markers, known as fiducial markers, in the environment. The sensors on the tracked device are designed to detect these specific markers.

  • Mechanism: Cameras or other optical sensors on the device identify the known patterns or shapes of the fiducial markers (for example square printed patterns, colored rings, or infrared (IR) LEDs). By analyzing the position, size, and orientation of these markers in the sensor's view, the system can calculate the device's pose relative to them. In 1999 Hirokazu Kato and Mark Billinghurst proposed a computer-vision method for tracking fiducial markers, and a calibration method for optical see-through HMDs based on that marker tracking, for a video-based AR conferencing system.[12] Welch and Foxlin cite earlier work at the University of North Carolina in which LEDs placed on physical objects, and later coin-size multicolored rings, served as landmarks for cameras mounted on an AR display.[1]
  • Limitations: Tracking only functions when the markers are within the sensor's field of view and are not occluded. It requires setting up the environment with markers in advance.[2] Valve's "Room" demo, shown to about 80 attendees of Steam Dev Days in January 2014, used fiducial markers placed on the walls and cameras on the headset; according to Road to VR's account of Valve's prototype timeline, it was obvious from the start that people would not put fiducials on their walls, and Valve went on to develop laser tracking, which became Lighthouse.[13]
  • Example: While not typically used for modern VR HMDs, the Nintendo Wii Remote uses a form of marker-based inside-out tracking. The remote contains an infrared camera, and the stationary Sensor Bar carries infrared lights that help the Wii Remote triangulate its position (primarily for pointing at the screen, not full 6DoF positional tracking).[14][15]

Markerless inside-out tracking

Main article: Markerless inside-out tracking

This is the most common approach in modern standalone VR and MR headsets. It uses computer vision techniques to track the device's position by recognizing natural features in the surrounding environment, without any artificial markers.

  • Mechanism: Cameras on the HMD capture video of the environment. Algorithms, usually based on SLAM (Simultaneous Localization and Mapping) or Visual-inertial odometry (VIO, also referred to as VIO), identify and track distinct features (corners, edges, textures) in the scene. As the headset moves, the system tracks how these features shift across the cameras' views to calculate the device's motion. Data from onboard Inertial Measurement Units (IMUs, containing accelerometers and gyroscopes) is fused with the visual data. The IMU provides high-frequency motion data and helps predict position during moments when visual tracking is temporarily lost (for example fast movements or poor lighting).[16] Apple's ARKit documentation gives a concise description of the visual-inertial method: ARKit recognizes notable features in the camera image, tracks how their positions change across video frames, and compares that information with motion-sensing data to model the device's position and motion.[17] Google's ARCore likewise describes its motion tracking as SLAM that detects visually distinct "feature points" and combines them with inertial measurements from the phone's IMU.[18] On Microsoft HoloLens 2, four grayscale visible-light tracking cameras running at 30 frames per second feed the system's real-time visual-inertial SLAM.[19]
  • Environment mapping: SLAM-based systems build and update a map of the environment as the user moves around, which lets the device re-localize itself within a known space. Meta said in 2018 that Oculus Insight detects thousands of points in the environment to compute an accurate position of the headset every millisecond.[20] The system generates a real-time 3D map of the user's immediate surroundings and computes the headset's position within that map; by tracking features such as the corners of a couch or the edge of a table, it can triangulate the user's location in the room.[4] Insight can also store a "spatial map" of an environment on the device and retrieve it later, and in a 2018 demo Meta shared one map between several headsets over a network so they could locate one another in the same room.[20] In the research literature, Andrew Davison and colleagues described their MonoSLAM system as achieving "drift-free performance" with a single camera by building a sparse but persistent map of natural landmarks.[21]

Controller tracking

Tracking the HMD is only part of the problem. Most VR systems also need to track handheld controllers, which is harder for an inside-out system because the controllers are separate objects that frequently leave the headset cameras' view. The common solution is optical-inertial: the controllers carry a ring or cluster of infrared LEDs that the headset's outward-facing cameras detect, while an IMU inside each controller supplies high-rate motion data.[4] Facebook engineering manager Andrew Melim described the Quest and Rift S method in 2019: the cameras capture images with extremely short exposure times, the software searches each frame for bright infrared "blobs", matches them to the LEDs on a 3D model of the controller, computes the controller's pose relative to the headset and fuses it with inertial data.[22] Solving the pose is a Perspective-n-Point problem; Melim wrote that three visible LEDs are enough in theory, but the system requires at least four correct matches to solve the pose robustly, with special solvers for two or three matches when a good prior pose is available.[23] The Quest controllers have 15 LEDs against 22 on the Rift's Touch controllers, and while a Rift controller was typically seen by two or three external cameras at once, there are very few poses in which more than one Quest camera sees a controller.[22] Matching is hardest when a controller is near the edge of the cameras' view, too close, too far or occluded, and losing it near the headset or the edges of the field of view can drop tracking to orientation-only (3DoF).[23]

Meta engineers noted that the infrared LEDs in the Quest controllers change appearance as they move closer to or farther from the headset, which the tracking algorithm has to account for; the team validated its tracking against OptiTrack motion-capture cameras during development.[4][24] Bright lights in the room are another source of error: in December 2019 Melim described how hundreds of holiday lights could be mistaken for controller LEDs, and how Meta added detection of stationary 3D lights and a small convolutional neural network that rejects blobs too large or too small to be controller LEDs.[25] PlayStation VR2 uses the same general scheme: four cameras on the headset track the headset and the controllers, and each VR2 Sense controller carries a tracking ring with IR LEDs that the headset cameras detect.[5]

Later designs moved away from the tracking ring. The Meta Quest Touch Pro Controllers, introduced with the Meta Quest Pro in October 2022, are self-tracking, with three cameras and an onboard chipset, so tracking works at any angle regardless of where the headset is facing.[26] The Meta Quest Touch Plus Controllers of the Meta Quest 3 keep infrared LEDs but place them on the controller face instead of a ring; Meta CTO Andrew Bosworth explained that Quest 3 continuously runs its controller-free hand tracking and fuses it with the LED tracking, with the optical data combined with the accelerometers and gyroscopes.[27] On Magic Leap 2, the headset's world cameras, which provide head pose at 30 Hz, can switch to a short-exposure infrared mode to track IR-emitting objects such as the controller.[28] Self-tracking body trackers use the same idea: HTC's VIVE Ultimate Tracker uses two wide-field-of-view cameras and computer vision for 6DoF inside-out tracking, with no base stations needed.[29]

Hand tracking

Some headsets avoid controllers entirely and track the user's hands directly with the same outward-facing cameras. The Apple Vision Pro, which became available in the United States on February 2, 2024, uses an input system controlled by the user's eyes (eye tracking), hands and voice; Apple said its R1 chip processes input from 12 cameras, five sensors and six microphones.[6] Apple's specifications for the current model (with the M5 chip) list six world-facing tracking cameras, four eye-tracking cameras, a LiDAR Scanner and four IMUs among the sensors, and give the R1 a 12-millisecond photon-to-photon latency.[30] On the Oculus Quest, hand tracking used the headset's monochrome tracking cameras: the MEgATrack system by Facebook researchers (SIGGRAPH 2020) uses four fisheye monochrome cameras, runs at 60 Hz on a PC and 30 Hz on a mobile processor, and was the default hand-tracking feature on the Oculus Quest.[31] HoloLens 2 instead uses a short-range mode of its depth camera, running at 45 frames per second, for hand tracking.[19] An independent evaluation of Meta Quest 2 hand tracking published in Behavior Research Methods in 2023 measured an average fingertip positional error of 1.1 cm, an average finger joint angle error of 9.6 degrees and an average delay of 45.0 ms.[32]

Contrast with outside-in tracking

In outside-in tracking the sensing hardware is fixed in the room and the tracked device carries markers or receivers. The Constellation system used by the Oculus Rift detected infrared LEDs on the tracked devices in short-exposure images from cameras placed externally, for example on a desk or table; depending on the setup, a Rift controller was typically seen by two or three of these cameras at a time.[22] When Oculus announced the Rift S in March 2019, it replaced this outside-in Constellation system with five-camera Insight inside-out tracking.[9]

The geometry of the two approaches has different strengths. Welch and Foxlin note that outside-looking-in keeps the target-borne component simple (as small as a retroreflector), but that the target's orientation must then be estimated from the relative positions of several target-borne light sources, a relative measurement that is sensitive to the distance of the fixed sensors.[1] The 1990 UNC paper used this argument for its inside-out design: with LEDs on antennae with a 0.5 m baseline on the helmet, a 0.1 degree head rotation moved the LEDs by only about 0.4 mm as seen by external cameras, whereas a helmet camera viewing ceiling beacons sees a much larger shift of the LED images for the same rotation, so it needs less resolution.[3] All optical systems need a clear line of sight between source and sensor.[1]

Because the reference cameras never move, outside-in systems have historically offered very stable, low-latency tracking, but they require mounting and calibrating external hardware and they suffer when the tracked object is hidden from the fixed cameras (occlusion). Inside-out tracking reverses the geometry: the cameras move with the user, so there is no external hardware to install, but the headset must do all of the computer-vision work itself and controller occlusion becomes the dominant tracking failure mode rather than body occlusion.[2][4][23] Measurements do not always favor the external system. In a 2021 study in a 5 m by 5 m room, Valentin Holzwarth and colleagues compared an Oculus Quest 2 with SteamVR Tracking using base stations 2.0 and Vive Trackers, and found significantly higher accuracy for the Quest 2 in the height of a tracked object and substantially higher positional precision; they noted that latency and rooms with fewer landmarks or different lighting still needed study.[33]

Advantages

  • Simpler setup: No need to install external sensors, base stations, or markers in the room.[4][34]
  • Portability: Systems are self-contained, which makes them easier to move between locations.[10]
  • Large tracking volume: Tracking is limited by what the sensors can see, not by the placement of external hardware. At Oculus Connect 5 in 2018, Meta demonstrated an arena-scale prototype built on Oculus Insight in a 4,000 square-foot play area with no external sensors; the arena included design elements on the ceiling, floor and throughout the space so that Insight could track players anywhere.[20] HTC recommends a play space of up to 10 by 10 m for the inside-out tracking of the VIVE Focus 3.[35]
  • Lower system cost: Removing external tracking hardware can reduce the overall cost of a VR system, although it increases the complexity and computational load on the headset itself. Holzwarth and colleagues cited the Quest 2's lower acquisition cost, higher mobility and easier setup compared with SteamVR Tracking.[33]

Disadvantages

  • Controller occlusion: Controller tracking can be lost if a controller moves outside the field of view of the HMD's cameras (for example behind the user's back, or too close to the HMD). Predictive algorithms using IMU data mitigate this for short periods.[16][23][10]
  • Environmental dependence: Tracking quality can degrade in poor lighting, in spaces with few visual features (for example blank walls), or with highly repetitive textures. Apple's ARKit documentation states that tracking quality is reduced when the camera cannot see details, such as when it points at a blank wall or the scene is too dark, and that excessive motion blurs the image.[17] The Oculus Santa Cruz prototype was demonstrated in a brightly lit room with no windows or mirrors, so their effect on tracking could not be tested.[36] Strong point lights can also confuse LED-based controller tracking.[25]
  • Computational load: Running SLAM/VIO requires significant processing on the device itself, which can affect battery life or require more powerful onboard processors. Meta's Insight team had to make its tracking efficient enough to run on a battery-powered headset with a mobile chipset.[4]
  • Drift: Inertial tracking accumulates drift, and markerless tracking can accumulate small positional errors over time (drift), which may require occasional recalibration or re-localization, although environment mapping reduces this.[16][21]
  • Initialization: The system needs to initialize and map a new area before tracking becomes fully stable. ARKit's documentation notes that the system develops a better understanding of the scene when the device is moving, and that detected planes are refined over time.[17]

History and adoption

Research origins

Early inside-out optical trackers for HMDs were built at the University of North Carolina at Chapel Hill. Jih-fang Wang, Vernon Chi and Henry Fuchs presented a prototype in 1990 with three cameras on the user's helmet, each using a lateral-effect photodiode, that viewed a regular pattern of flashing infrared LEDs on the ceiling; it gave more than 200 updates per second and registered 0.1 degree rotations and 2 mm translations in a working volume of about 1,000 cubic feet.[3] A companion paper with Ronald Azuma, Gary Bishop and John Eyles described the plan for a full-scale version with ceiling panels holding the beacons, so that the user could walk almost anywhere under the grid.[37] UNC demonstrated a working scalable ceiling tracker at SIGGRAPH 1991, using four head-mounted sensors looking up at LEDs in machined ceiling panels. Its successor, the HiBall Tracker (1999), replaced the bulky cameras with a single head-mounted sensor cluster of about 300 grams and generated over 2,000 head-pose estimates per second with less than one millisecond of latency in a 4.5 by 8.5 m room.[38] Welch and Foxlin noted in 2002 that if walk-through VR systems came to be built on wearable computers, neither the HiBall nor InterSense's IS-900 would be really appropriate, and a wearable self-tracking system with no active infrastructure would be much more easily integrated.[1]

Markerless tracking came from robotics and computer vision. MonoSLAM (2007) was described by its authors as the first successful application of SLAM from mobile robotics to a single uncontrolled camera, running at 30 Hz and demonstrated for live augmented reality with a hand-held camera.[21] The same year, Georg Klein and David Murray's PTAM split tracking and mapping into parallel threads to track a hand-held camera in small AR workspaces with maps of thousands of landmarks,[39] and Anastasios Mourikis and Stergios Roumeliotis published the multi-state constraint Kalman filter, an EKF-based method for vision-aided inertial navigation whose cost grows only linearly with the number of features.[40] Game Developer noted in 2019 that SLAM had previously been used in robotics and in smartphone AR camera effects to map out shapes in a space.[24] Oculus Insight's lead machine perception architect, Oskar Linde, had co-founded 13th Lab, which Meta (then Facebook) says demonstrated the first use of visual SLAM in a consumer application in a 2011 mobile AR game; Facebook acquired 13th Lab in 2014, and many of the Zurich engineers on the Insight team came from Zurich Eye, a program of ETH Zurich and the University of Zurich.[4]

Consumer devices

Microsoft's first HoloLens AR headset had an 11-element sensor cluster, and Road to VR's Ben Lang called its inside-out tracking "rock solid" in 2016, noting that it held virtual objects in place during fast head movements and ignored people walking around on a busy show floor.[41] Oculus (then owned by Facebook) publicly demonstrated inside-out tracking for VR with the "Santa Cruz" prototype at Oculus Connect 3 in October 2016, describing it as a feature prototype that opened a path to standalone VR; the prototype was a modified Oculus Rift with on-board compute, whose four cameras and IMU handled all tracking.[36][42] Lang reported that Oculus said the demo used no pre-mapping or hidden fiducial markers.[36]

The Windows Mixed Reality (WMR) headsets tracked both the headset and the controllers with cameras on the headset, and PCWorld's reviewer called them essentially the first mainstream attempt at camera-driven inside-out tracking.[10] Acer, Dell, HP and Lenovo launched the first WMR headsets on October 17, 2017, at prices beginning at US$399 with motion controllers, and Microsoft said Samsung's HMD Odyssey would follow on November 6.[34] WMR's controllers were criticized for tracking quality compared with contemporary Rift and Vive controllers, because they were tracked only by the two cameras on the front of the headset, so a controller lost positional tracking whenever it moved out of that camera view.[10] Microsoft deprecated Windows Mixed Reality in December 2023 and removed it from Windows 11 with version 24H2 (the Windows 11 2024 Update), released at the start of October 2024, which left the headsets unsupported on that version and later.[43] Google's Daydream standalone platform, whose first headset was the Lenovo Mirage Solo, used WorldSense 6DoF inside-out tracking.[11]

The technology reached the mass market with Meta's Oculus Insight, which shipped in the Oculus Quest and Oculus Rift S in 2019. Insight uses cameras on the headset (four ultra-wide-angle cameras on Quest, five cameras on Rift S) plus IMU data and on-device SLAM to track the headset and two controllers simultaneously; Meta described the tracking as accurate to the millimeter, running on the Quest's mobile chipset.[4][20][9] Since then, inside-out tracking has become the default for standalone headsets, including the Meta Quest 2, Quest 3, Quest 3S and Quest Pro, the HTC Vive Focus series and HTC Vive XR Elite, the Pico line and the Apple Vision Pro, and it is also used by the console-tethered PlayStation VR2.[35][44][45][5][6] AR headsets including Microsoft HoloLens 1 and 2 and Magic Leap 1 and 2 also use inside-out tracking.[41][19][28]

Recent headsets continue the camera-based approach. Samsung's Galaxy XR, the first headset built on Android XR, went on sale in the United States on October 21, 2025 for US$1,799.99; its sensors include six world-facing tracking cameras, two pass-through cameras, four eye-tracking cameras, five IMUs and a depth sensor, and Samsung says it tracks the user's head, hands and eyes.[7] Valve's earlier Lighthouse system, used by the HTC Vive, depends on base stations in the room;[10] the Steam Frame instead uses camera-based tracking. Valve states that four high-resolution monochrome cameras provide controller and headset tracking with no setup required, and that infrared illuminators on the headset keep tracking working in dark environments.[8] Valve opened Steam Frame purchase reservations in September 2026.[46]

Examples of devices and platforms using inside-out tracking

Many modern VR, AR, and MR systems use markerless inside-out tracking:

The table lists the inside-out tracking hardware of representative systems as stated by their makers or in the cited reporting.

Device or platform Type Inside-out tracking hardware Source
Oculus Santa Cruz VR prototype Four cameras and an IMU on a modified Rift [36]
Microsoft HoloLens AR headset 11-element sensor cluster [41]
First-generation Windows Mixed Reality headsets PC VR headsets Two front cameras [10]
Lenovo Mirage Solo Standalone VR (Daydream) WorldSense 6DoF tracking [11]
Oculus Quest Standalone VR Four ultra-wide-angle cameras (Oculus Insight) [20]
Oculus Rift S PC VR Five cameras (Oculus Insight) [9]
HTC Vive Cosmos PC VR Six cameras with the default faceplate (two of them on the removable faceplate) [47]
Microsoft HoloLens 2 AR headset Four grayscale visible-light tracking cameras (30 fps) for visual-inertial SLAM; depth camera for hand tracking [19]
HTC Vive Focus 3 Standalone VR Four tracking cameras [35]
HTC Vive XR Elite Standalone VR/MR Four tracking cameras and a depth sensor [44]
Meta Quest Touch Pro Controllers Self-tracking controllers Three cameras and an onboard chipset [26]
Magic Leap 2 AR headset Three world cameras (head pose at 30 Hz) and a depth camera [28]
PlayStation VR2 Console VR Four cameras for headset and controller tracking [5]
VIVE Ultimate Tracker Body tracker Two wide-field-of-view cameras [29]
Apple Vision Pro MR headset 12 cameras and five sensors (original model); six world-facing tracking cameras listed for the M5 model [6][30]
Pico 4 Ultra Standalone VR/MR Four environment tracking cameras and an iToF depth camera [45]
Samsung Galaxy XR MR headset (Android XR) Six world-facing tracking cameras, five IMUs and a depth sensor [7]
Steam Frame VR headset Four monochrome cameras with infrared illuminators [8]

See also

References

  1. ↑ 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 Greg Welch, Eric Foxlin (2002-11). "Motion Tracking: No Silver Bullet, but a Respectable Arsenal". IEEE Computer Graphics and Applications, vol. 22, no. 6, pp. 24-38. https://doi.org/10.1109/MCG.2002.1046626. Retrieved 2026-10-17.
  2. ↑ 2.0 2.1 2.2 Boger, Yuval (2014-08-04). "Positional tracking: "Outside-in" vs. "Inside-out"". The VR Guy (vrguy.blogspot.com). http://vrguy.blogspot.com/2014/08/positional-tracking-outside-in-vs.html. Retrieved 2026-10-17.
  3. ↑ 3.0 3.1 3.2 Jih-fang Wang, Vernon Chi, Henry Fuchs (1990). "A Real-time Optical 3D Tracker for Head-mounted Display Systems". Proceedings of the 1990 Symposium on Interactive 3D Graphics (Computer Graphics, vol. 24, no. 2), pp. 205-215. ACM. doi:10.1145/91385.91447. https://techreports.cs.unc.edu/papers/90-011.pdf. Retrieved 2026-10-17.
  4. ↑ 4.00 4.01 4.02 4.03 4.04 4.05 4.06 4.07 4.08 4.09 4.10 "From the lab to the living room: The story behind Facebook's Oculus Insight technology and a new era of consumer VR". Tech at Meta. Meta. 2019-08-21. https://tech.facebook.com/reality-labs/2019/8/the-story-behind-oculus-insight-technology/. Retrieved 2026-10-17.
  5. ↑ 5.0 5.1 5.2 5.3 Sid Shuman (2023-02-06). "PlayStation VR2 - The ultimate FAQ". PlayStation.Blog. https://blog.playstation.com/2023/02/06/playstation-vr2-the-ultimate-faq/. Retrieved 2026-10-17.
  6. ↑ 6.0 6.1 6.2 6.3 "Apple Vision Pro available in the U.S. on February 2". Apple Newsroom. Apple. 2024-01-08. https://www.apple.com/newsroom/2024/01/apple-vision-pro-available-in-the-us-on-february-2/. Retrieved 2026-10-17.
  7. ↑ 7.0 7.1 7.2 "Introducing Galaxy XR: Opening New Worlds". Samsung Newsroom U.S.. Samsung Electronics. 2025-10-21. https://news.samsung.com/us/introducing-galaxy-xr-opening-new-worlds/. Retrieved 2026-10-17.
  8. ↑ 8.0 8.1 8.2 "Steam Frame". Steam. Valve. https://store.steampowered.com/hardware/steamframe. Retrieved 2026-10-17.
  9. ↑ 9.0 9.1 9.2 9.3 Ben Lang (2019-03-20). "Oculus Rift S Revealed with Inside-out Tracking, Resolution Bump, & New Ergonomics". Road to VR. https://roadtovr.com/oculus-rift-s-specs-release-date-announcement-gdc-2019/. Retrieved 2026-10-17.
  10. ↑ 10.0 10.1 10.2 10.3 10.4 10.5 10.6 10.7 Hayden Dingman (2018-05-11). "Microsoft Windows Mixed Reality review: Easy to set up, hard to use". PCWorld. Archived from the original on 2021-09-24. https://web.archive.org/web/20210924184610/https://www.pcworld.com/article/3269791/windows-mixed-reality-review-steamvr.html. Retrieved 2026-10-17.
  11. ↑ 11.0 11.1 11.2 "Daydream Standalone". Google VR, Google for Developers. Google. https://developers.google.com/vr/discover/daydream-standalone. Retrieved 2026-10-17.
  12. ↑ Hirokazu Kato, Mark Billinghurst (1999). "Marker tracking and HMD calibration for a video-based augmented reality conferencing system". Proceedings of the 2nd IEEE and ACM International Workshop on Augmented Reality (IWAR '99), pp. 85-94. IEEE. https://doi.org/10.1109/IWAR.1999.803809. Retrieved 2026-10-17.
  13. ↑ Paul James (2015-03-05). "Valve Reveals Timeline of Vive Prototypes, We Chart it For You". Road to VR. https://roadtovr.com/valve-reveals-timeline-of-vive-prototypes-we-chart-it-for-you/. Retrieved 2026-10-17.
  14. ↑ Johnny Chung Lee (2008-02-13). "Interaction Techniques Using the Wii Remote". Stanford EE Computer Systems Colloquium (EE380). Stanford University. https://web.stanford.edu/class/ee380/Abstracts/080213.html. Retrieved 2026-10-17.
  15. ↑ "How to Check Functionality of the Sensor Bar (Wii U)". Nintendo Support. Nintendo of America. https://en-americas-support.nintendo.com/app/answers/detail/a_id/1155. Retrieved 2026-10-17.
  16. ↑ 16.0 16.1 16.2 Boger, Yuval (2014-06-02). "Overview of Positional Tracking Technologies for Virtual Reality". Road to VR. http://www.roadtovr.com/overview-of-positional-tracking-technologies-virtual-reality/. Retrieved 2026-10-17.
  17. ↑ 17.0 17.1 17.2 "Understanding World Tracking". Apple Developer Documentation. Apple. https://developer.apple.com/documentation/arkit/understanding-world-tracking. Retrieved 2026-10-17.
  18. ↑ "Fundamental concepts". ARCore, Google for Developers. Google. https://developers.google.com/ar/develop/fundamentals. Retrieved 2026-10-17.
  19. ↑ 19.0 19.1 19.2 19.3 Dorin Ungureanu, Federica Bogo, Silvano Galliani, Pooja Sama, Xin Duan, Casey Meekhof, Jan Stühmer, Thomas J. Cashman, Bugra Tekin, Johannes L. Schönberger, Pawel Olszta, Marc Pollefeys (2020-08-25). "HoloLens 2 Research Mode as a Tool for Computer Vision Research". arXiv (Microsoft technical report). arXiv:2008.11239. https://arxiv.org/abs/2008.11239. Retrieved 2026-10-17.
  20. ↑ 20.0 20.1 20.2 20.3 20.4 "Playing with the future: Oculus Insight powers beyond-room-scale VR gaming". Tech at Meta. Meta. 2018-09-26. Archived from the original on 2026-06-08. https://web.archive.org/web/20260608142700/https://tech.facebook.com/reality-labs/2018/9/playing-with-the-future-oculus-insight-powers-beyond-room-scale-vr-gaming/. Retrieved 2026-10-17.
  21. ↑ 21.0 21.1 21.2 Andrew J. Davison, Ian D. Reid, Nicholas D. Molton, Olivier Stasse (2007-06). "MonoSLAM: Real-Time Single Camera SLAM". IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 29, no. 6, pp. 1052-1067. https://doi.org/10.1109/TPAMI.2007.1049. Retrieved 2026-10-17.
  22. ↑ 22.0 22.1 22.2 Andrew Melim (2019-09-20). "Increasing Fidelity with Constellation-Tracked Controllers". Meta Horizon OS Developers Blog. Meta. https://developers.meta.com/horizon/blog/increasing-fidelity-with-constellation-tracked-controllers/. Retrieved 2026-10-17.
  23. ↑ 23.0 23.1 23.2 23.3 Andrew Melim (2019-11-04). "Tracking Technology Explained: LED Matching". Meta Horizon OS Developers Blog. Meta. https://developers.meta.com/horizon/blog/tracking-technology-explained-led-matching/. Retrieved 2026-10-17.
  24. ↑ 24.0 24.1 Alissa McAloon (2019-08-22). "C'mon and SLAM - How Oculus tackled portable, 6DOF tracking for the Quest". Game Developer. https://www.gamedeveloper.com/game-platforms/c-mon-and-slam-how-oculus-tackled-portable-6dof-tracking-for-the-quest. Retrieved 2026-10-17.
  25. ↑ 25.0 25.1 Andrew Melim (2019-12-20). "Optimizing Oculus Insight Controller Tracking to Work in Challenging Conditions (Like Near Holiday Lights)". Meta Horizon OS Developers Blog. Meta. https://developers.meta.com/horizon/blog/optimizing-oculus-insight-controller-tracking-to-work-in-challenging-conditions-like-near-holiday-lights/. Retrieved 2026-10-17.
  26. ↑ 26.0 26.1 David Heaney (2022-10-11). "Quest Pro's Controllers Support Quest 2 For $300 Later This Year". UploadVR. https://www.uploadvr.com/quest-touch-pro-controllers-quest-2/. Retrieved 2026-10-17.
  27. ↑ David Heaney (2023-06-03). "Meta Reveals How Quest 3 Controller Tracking Works As Beat Saber Founder Gives His Take". UploadVR. https://www.uploadvr.com/meta-explains-quest-3-controller-tracking/. Retrieved 2026-10-17.
  28. ↑ 28.0 28.1 28.2 "Sensors". Magic Leap Developer Documentation. Magic Leap. https://developer-docs.magicleap.cloud/docs/device/hardware/sensors/. Retrieved 2026-10-17.
  29. ↑ 29.0 29.1 "VIVE Ultimate Tracker". VIVE. HTC. https://www.vive.com/us/accessory/vive-ultimate-tracker/. Retrieved 2026-10-17.
  30. ↑ 30.0 30.1 "Apple Vision Pro - Technical Specifications". Apple. Apple. https://www.apple.com/apple-vision-pro/specs/. Retrieved 2026-10-17.
  31. ↑ Shangchen Han, Beibei Liu, Randi Cabezas, Christopher D. Twigg, Peizhao Zhang, Jeff Petkau, Tsz-Ho Yu, Chun-Jung Tai, Muzaffer Akbay, Zheng Wang, Asaf Nitzan, Gang Dong, Yuting Ye, Lingling Tao, Chengde Wan, Robert Wang (2020). "MEgATrack: Monochrome Egocentric Articulated Hand-Tracking for Virtual Reality". ACM Transactions on Graphics, vol. 39, no. 4 (SIGGRAPH 2020). https://doi.org/10.1145/3386569.3392452. Retrieved 2026-10-17.
  32. ↑ Diar Abdlkarim, Massimiliano Di Luca, Poppy Aves, Mohamed Maaroufi, Sang-Hoon Yeo, R. Chris Miall, Peter Holland, Joseph M. Galea (2023-02-13). "A methodological framework to assess the accuracy of virtual reality hand-tracking systems: A case study with the Meta Quest 2". Behavior Research Methods, vol. 56, no. 2, pp. 1052-1063. https://doi.org/10.3758/s13428-022-02051-8. Retrieved 2026-10-17.
  33. ↑ 33.0 33.1 Valentin Holzwarth, Joy Gisler, Christian Hirt, Andreas Kunz (2021-03-20). "Comparing the Accuracy and Precision of SteamVR Tracking 2.0 and Oculus Quest 2 in a Room Scale Setup". Proceedings of the 2021 5th International Conference on Virtual and Augmented Reality Simulations (ICVARS 2021), pp. 42-46. ACM. https://doi.org/10.1145/3463914.3463921. Retrieved 2026-10-17.
  34. ↑ 34.0 34.1 Yusuf Mehdi (2017-10-17). "Windows 10 Fall Creators Update and mixed reality headsets available today; Announcing Surface Book 2". Windows Experience Blog. Microsoft. https://blogs.windows.com/windowsexperience/2017/10/17/windows-10-fall-creators-update-and-mixed-reality-headsets-available-today-announcing-surface-book-2/. Retrieved 2026-10-17.
  35. ↑ 35.0 35.1 35.2 "VIVE Focus 3 Specs". VIVE. HTC. https://www.vive.com/us/product/vive-focus3/specs/. Retrieved 2026-10-17.
  36. ↑ 36.0 36.1 36.2 36.3 Ben Lang (2016-10-06). "Hands-on - Oculus' Wireless 'Santa Cruz' Prototype Makes Standalone Room-scale Tracking a Reality". Road to VR. https://www.roadtovr.com/hands-on-oculus-wireless-santa-cruz-prototype-makes-standalone-room-scale-tracking-a-reality/. Retrieved 2026-10-17.
  37. ↑ Jih-fang Wang, Ronald Azuma, Gary Bishop, Vernon Chi, John Eyles, Henry Fuchs (1990). "Tracking a head-mounted display in a room-sized environment with head-mounted cameras". Proceedings of SPIE, vol. 1290, pp. 47-57. SPIE. doi:10.1117/12.20954. https://www.ronaldazuma.com/papers/orlando.pdf. Retrieved 2026-10-17.
  38. ↑ Greg Welch, Gary Bishop, Leandra Vicci, Stephen Brumback, Kurtis Keller, D'nardo Colucci (1999-12). "The HiBall Tracker: High-Performance Wide-Area Tracking for Virtual and Augmented Environments". Proceedings of the ACM Symposium on Virtual Reality Software and Technology (VRST '99). ACM. https://doi.org/10.1145/323663.323664. Retrieved 2026-10-17.
  39. ↑ Georg Klein, David Murray (2007-11). "Parallel Tracking and Mapping for Small AR Workspaces". 2007 6th IEEE and ACM International Symposium on Mixed and Augmented Reality (ISMAR), pp. 1-10. IEEE. https://doi.org/10.1109/ISMAR.2007.4538852. Retrieved 2026-10-17.
  40. ↑ Anastasios I. Mourikis, Stergios I. Roumeliotis (2007-04). "A Multi-State Constraint Kalman Filter for Vision-aided Inertial Navigation". Proceedings 2007 IEEE International Conference on Robotics and Automation (ICRA), pp. 3565-3572. IEEE. https://doi.org/10.1109/ROBOT.2007.364024. Retrieved 2026-10-17.
  41. ↑ 41.0 41.1 41.2 Ben Lang (2016-09-20). "Hololens' Inside-out Tracking Is Game Changing for AR & VR, and No One Is Talking about It". Road to VR. https://www.roadtovr.com/microsoft-hololens-inside-out-tracking-augmented-reality-virtual-reality-ar-vr/. Retrieved 2026-10-17.
  42. ↑ "Oculus Reveals Touch Launch Details and Unveils Santa Cruz Prototype at Oculus Connect 3". Meta Newsroom. Meta. 2016-10-06. https://about.fb.com/news/2016/10/oculus-reveals-touch-launch-details-and-unveils-santa-cruz-prototype-at-oculus-connect-3/. Retrieved 2026-10-17.
  43. ↑ Scott Hayden (2024-10-02). "Windows 11 No Longer Supports Microsoft's Windows VR Headsets Following October Update". Road to VR. https://roadtovr.com/windows-11-drops-wmr-support-24h2/. Retrieved 2026-10-17.
  44. ↑ 44.0 44.1 "VIVE XR Elite Specs". VIVE. HTC. https://www.vive.com/us/product/vive-xr-elite/specs/. Retrieved 2026-10-17.
  45. ↑ 45.0 45.1 "PICO 4 Ultra Specs". PICO. PICO (ByteDance). https://www.picoxr.com/global/products/pico4-ultra/specs. Retrieved 2026-10-17.
  46. ↑ Scott Hayden (2026-09-14). "Valve Announces Steam Frame Price, Release Date, and Accessories - Pre-orders Now Open". Road to VR. https://roadtovr.com/valve-steam-frame-price-release-pre-orders/. Retrieved 2026-10-17.
  47. ↑ David Heaney (2019-06-21). "HTC Vive Cosmos Faceplate Update Brings Tracking Cameras To 6". UploadVR. https://www.uploadvr.com/vive-cosmos-6-cameras-faceplate/. Retrieved 2026-10-17.
Retrieved from "https://vrarwiki.com/index.php?title=Inside-out_tracking&oldid=52426"