Jump to content

IMU: Difference between revisions

From VR & AR Wiki
← Older editNewer edit →
VisualWikitext
No edit summary
No edit summary
Tag: Reverted
Line 1: Line 1:
{{see also|Terms|Technical Terms}}
{{See also|Terms|Technical terms}}
An '''Inertial Measurement Unit''' ('''IMU''') is an electronic [[sensor]] [[device]] that measures and reports a body's specific force, angular rate, and sometimes the orientation of the body, using a combination of [[accelerometer]]s, [[gyroscope]]s, and often [[magnetometer]]s.<ref name="TDK_IMU_Overview"> TDK InvenSense. What is an Inertial Measurement Unit (IMU)? [https://invensense.tdk.com/technology/motion/imu/ TDK InvenSense Website]. Accessed October 26, 2023.</ref> IMUs are fundamental components in [[virtual reality|Virtual Reality (VR)]] and [[augmented reality|Augmented Reality (AR)]] systems for tracking the orientation of [[Head-Mounted Display|HMDs]] and [[Input Devices]] like controllers.


==Components and Function==
An '''Inertial Measurement Unit''' ('''IMU''') is an electronic [[sensor]] [[device]] that measures and reports a body's specific force, angular rate, and sometimes orientation using a combination of [[accelerometer]]s, [[gyroscope]]s, and often [[magnetometer]]s.<ref name="TDK_IMU_Overview">“MEMS Motion Sensors – Inertial Measurement Units.” TDK InvenSense. https://invensense.tdk.com/ (accessed 2 May 2025).</ref> IMUs are fundamental components in [[virtual reality|Virtual Reality (VR)]] and [[augmented reality|Augmented Reality (AR)]] systems, tracking the orientation of [[Head‑Mounted Display|HMDs]] and [[Input Devices]] like controllers.
 
== Components and Function ==
A typical IMU integrates multiple sensor types onto a microchip:
A typical IMU integrates multiple sensor types onto a microchip:
* '''[[Accelerometer]]s''': Measure proper acceleration (g‑force), including gravity.<ref name="Woodman_IMU_Tutorial">Woodman, O. J. (2007) ''An Introduction to Inertial Navigation''. Technical Report UCAM‑CL‑TR‑696, University of Cambridge Computer Laboratory. https://www.cl.cam.ac.uk/techreports/UCAM-CL-TR-696.pdf</ref>
* '''[[Gyroscope]]s''': Measure [[angular velocity]] about one or more axes.<ref name="Woodman_IMU_Tutorial" />
* '''[[Magnetometer]]s''' (optional): Sense the local [[magnetic field]], providing an absolute yaw reference.<ref name="Woodman_IMU_Tutorial" />


*  '''[[Accelerometer]]s''': Measure proper acceleration (g-force), which includes both acceleration due to movement and the constant pull of [[gravity]].<ref name="Woodman_IMU_Tutorial"> Woodman, O. J. (2007). An introduction to inertial navigation. University of Cambridge Computer Laboratory Technical Report, UCAM-CL-TR-696. [https://www.cl.cam.ac.uk/techreports/UCAM-CL-TR-696.pdf PDF Link]</ref> When the IMU is relatively static, accelerometers can determine tilt (pitch and roll angles) relative to the direction of gravity. When moving, they measure linear acceleration.
An IMU containing all three sensors is called a 9‑axis IMU or a [[MARG]] (Magnetic, Angular‑Rate, and Gravity) sensor.<ref>Madgwick, S. O. H.; Harrison, A. J. L.; Vaidyanathan, R. (2011) “Estimation of IMU and MARG Orientation Using a Gradient Descent Algorithm.” ''Proc. IEEE ICORR'', 1‑7. https://vigir.missouri.edu/~gdesouza/Research/Conference_CDs/RehabWeekZürich/icorr/papers/Madgwick_Estimation_of_IMU_and_MARG_orientation_using_a_gradient_descent_algorithm_ICORR2011.pdf</ref>
*  '''[[Gyroscope]]s''': Measure [[angular velocity]] (rate of rotation) around one or more axes.<ref name="Woodman_IMU_Tutorial"/> In VR/AR, they detect rotational movements corresponding to [[pitch]] (nodding 'yes'), [[yaw]] (shaking 'no'), and [[roll]] (tilting head side-to-side). Gyroscopes provide fast and responsive rotational data but are prone to [[sensor drift]] over time.
*  '''[[Magnetometer]]s''' (Optional but common): Measure the strength and direction of the local [[magnetic field]], typically the Earth's magnetic field. They act like a compass to provide an absolute reference for the yaw orientation, helping to correct gyroscope drift around the vertical axis.<ref name="Woodman_IMU_Tutorial"/> However, they are susceptible to interference from nearby magnetic materials or electronic devices.
 
When an IMU includes all three sensors (accelerometer, gyroscope, and magnetometer), it is sometimes referred to as a 9-axis IMU or a [[MARG]] (Magnetic, Angular Rate, and Gravity) sensor.<ref>Madgwick, Sebastian OH, Andrew JL Harrison, and Ravi Vaidyanathan. "Estimation of IMU and MARG orientation using a gradient descent algorithm." IEEE international conference on rehabilitation robotics. IEEE, 2011.</ref>


==Sensor Fusion==
== Sensor Fusion ==
Raw data from individual sensors can be noisy (for example accelerometers during fast movement) and inaccurate (for example gyroscopes drift). [[Sensor fusion]] algorithms, such as [[Kalman filter]]s or complementary filters, are essential.<ref name="Mahony_Filter"> Mahony, R., Hamel, T., & Pflimlin, J. M. (2008). Nonlinear Complementary Filters on the Special Orthogonal Group. IEEE Transactions on Automatic Control, 53(5), 1203-1218. [https://ieeexplore.ieee.org/document/4532514 Abstract Link]</ref> These algorithms intelligently combine the data from the accelerometers, gyroscopes (and magnetometers, if present) to produce a single, more accurate, stable, and low-latency estimate of the device's orientation in real-time.
Raw data are noisy or drift; [[Sensor fusion]] algorithms such as complementary or [[Kalman filter]]s combine the signals to yield a stable estimate.<ref name="Mahony_Filter">Mahony, R.; Hamel, T.; Pflimlin, J.‑M. (2008) “Nonlinear Complementary Filters on the Special Orthogonal Group.” ''IEEE T‑AC'' 53 (5): 1203‑1218. https://doi.org/10.1109/TAC.2008.923738</ref>


== Role in VR/AR ==
== Role in VR/AR ==
IMUs are crucial for providing low-latency [[rotational tracking]], which is essential for creating a sense of [[immersion]] and preventing [[motion sickness]].<ref name="LaValle_VR_Book"> LaValle, S. M. (2016). Virtual Reality. Cambridge University Press. Chapter 9: Tracking. [http://lavalle.pl/vr/book.html Online Book Link]</ref>
IMUs provide the sub‑millisecond [[rotational tracking]] needed for [[immersion]] and for avoiding [[motion sickness]].<ref name="LaValle_VR_Book">LaValle, S. M. (2016) ''Virtual Reality'' – Chapter 9. https://lavalle.pl/vr/vrch9.pdf</ref>
 
=== Importance in Head Tracking ===
IMUs provide the rapid [[orientation tracking]] needed to update the virtual view in sync with the user's head movements. This low latency is critical for user comfort. The typical update rate of modern IMUs used in VR headsets is between 500Hz to 1000Hz, much faster than most visual tracking systems can achieve alone.<ref>Niehorster, Diederick C., Li Li, and Markus Lappe. "The accuracy and precision of position and orientation tracking in the HTC Vive virtual reality system for scientific research." i-Perception 8.3 (2017).</ref>
 
===3 Degrees of Freedom (DoF)===
An IMU inherently provides [[Degrees of Freedom|3 DoF tracking]], measuring orientation changes (pitch, yaw, roll). This is sufficient for basic VR experiences like 360-degree video viewing on mobile VR headsets where the user's physical position in the room is not tracked.


===6DoF Tracking Systems===
=== 6DoF Tracking Systems ===
For full [[6DoF]] tracking (which includes [[positional tracking]] translation along X, Y, and Z axes), IMU data is combined via sensor fusion with data from other tracking systems. These can include:
For full [[6DoF]] tracking, IMU data are fused with camera‑based [[inside‑out tracking]] or external [[lighthouse tracking]] systems to correct long‑term drift.<ref>Ercan, A. O.; Erdem, A. T. (2011) “On Sensor Fusion for Head Tracking in Augmented Reality Applications.” ''Proc. ACC'', 1286‑1291. https://doi.org/10.1109/ACC.2011.5991077</ref>
*  [[Inside-out tracking]]: Cameras on the HMD observe the external environment.
*  [[Outside-in tracking]]: External sensors (like cameras or [[lighthouse tracking|base stations]]) observe markers on the HMD and controllers.
*  [[Camera-based tracking]]: General term encompassing various visual tracking methods.
In these systems, the IMU provides the high-frequency orientation updates, while the positional tracking system provides absolute position data and periodically corrects for any accumulated IMU drift.<ref name="LaValle_VR_Book"/><ref>Hyvärinen, Timo, et al. "Sensor fusion for head tracking in augmented reality applications." 2019 IEEE International Symposium on Mixed and Augmented Reality Adjunct (ISMAR-Adjunct). IEEE, 2019.</ref>


== Limitations and Correction ==
== Limitations and Correction ==
While essential, IMUs have inherent limitations:
* '''Sensor drift''' from gyroscopes 
 
* '''Magnetic interference'''
*   '''[[Sensor Drift]]''': Gyroscopes accumulate small errors over time, leading to a gradual mismatch between the tracked orientation and the real-world orientation. This is particularly noticeable in yaw if uncorrected.
* '''No positional data''' without external references 
*   '''Magnetic Interference''': Magnetometers can be disturbed by ferrous materials or strong magnetic fields in the environment, leading to inaccurate yaw readings. Advanced sensor fusion algorithms may attempt to detect and compensate for such interference.
Drift is mitigated by visual corrections, complementary filters, and periodic resets such as [[Zero‑velocity update]]s (ZUPTs).<ref>Cadena, C. et al. (2016) “Past, Present, and Future of SLAM.” arXiv:1606.05830.</ref>
*   '''No Positional Data''': By themselves, IMUs cannot determine a device's position in space; they only measure rotation and linear acceleration, not absolute location or translational velocity relative to the world.
 
VR/AR systems address these limitations, particularly drift, through:
*  Visual correction using cameras or external reference points (in 6DoF systems)
*  [[Complementary filtering]] combining accelerometer (gravity vector) and gyroscope data for tilt correction
*  [[Kalman filtering]] algorithms integrating multiple sensor inputs and predictive models
*  Magnetometer data (if available and reliable) for absolute yaw correction
*  [[Zero velocity updates]] (ZUPTs) during periods of detected stillness to reset velocity error accumulation<ref>Cadena, Cesar, et al. "Past, present, and future of simultaneous localization and mapping: Toward the robust-perception age." IEEE Transactions on robotics 32.6 (2016): 1309-1332.</ref>
 
==IMU Specifications for VR/AR==
For optimal performance in VR/AR applications, IMUs typically require:
 
*  Low latency (< 2ms sensor processing time desirable)
*  High update rate (500-1000Hz)
*  High precision gyroscopes (< 0.01 degrees/second drift)
*  Low noise accelerometers
*  Efficient power consumption
*  Small form factor
*  Integrated processing capabilities (sometimes including basic sensor fusion)<ref>Angelini, Lee, et al. "Understanding sensors: prioritizations for selecting sensors in mobile VR applications." Internet Research (2022).</ref>


==Future Developments==
== IMU Specifications for VR/AR ==
Next-generation IMUs for VR/AR are focusing on:
* Latency &lt; 2 ms 
* Update rate 500–1000 Hz 
* Gyro drift &lt; 0.01 °/s 
* Low‑noise accelerometers 
* Integrated processing / basic fusion<ref name="TDK_Homepage">“Motion Sensors – Product Portfolio.” TDK InvenSense. https://invensense.tdk.com/products/motion-tracking/ (accessed 2 May 2025).</ref>


*  Reduced power consumption for longer device battery life
== Future Developments ==
*  Smaller form factors for integration into lighter HMDs and glasses
Trends include lower power draw, on‑sensor machine‑learning, and reduced drift.<ref>Yole Group (2024) ''Status of the MEMS Industry 2024'' (report brochure). https://www.yolegroup.com/product/report/status-of-the-mems-industry-2024/</ref>
*  Integrated [[machine learning|ML]] capabilities for improved motion prediction and pattern recognition
*  Enhanced sensor fusion algorithms, potentially running on the sensor itself
*  Further reduction in sensor noise and drift characteristics<ref>Adams, Michael D. "MEMS IMU Navigation with Model Based Dead-Reckoning and One-Way-Travel-Time Acoustic Measurements." IEEE Journal of Oceanic Engineering (2023).</ref>


==Key IMU Manufacturers==
== Key IMU Manufacturers ==
Several companies manufacture IMUs used in consumer electronics, including VR/AR devices:
* [[TDK]]‑[[InvenSense]]<ref name="TDK_Homepage" />
*   [[TDK]] [[Invensense]]<ref name="TDK_Homepage"> TDK InvenSense - Motion Sensors [https://invensense.tdk.com/products/motion-tracking/ TDK InvenSense Website]. Accessed October 26, 2023.</ref> - Major provider for consumer electronics.
* [[Bosch Sensortec]]<ref name="Bosch_Homepage">“Inertial Measurement Units (IMUs).” Bosch Sensortec. https://www.bosch-sensortec.com/products/motion-sensors/imus/</ref>
*   [[Bosch Sensortec]]<ref name="Bosch_Homepage"> Bosch Sensortec - IMUs [https://www.bosch-sensortec.com/products/motion-sensors/imus/ Bosch Sensortec Website]. Accessed October 26, 2023.</ref> - Produces high-performance MEMS sensors.
* [[STMicroelectronics]]<ref name="ST_Homepage">“MEMS and Sensors.” STMicroelectronics. https://www.st.com/en/mems-and-sensors.html</ref>
*   [[STMicroelectronics]]<ref name="ST_Homepage"> STMicroelectronics - MEMS Motion Sensors [https://www.st.com/en/mems-and-sensors/mems-motion-sensors.html STMicroelectronics Website]. Accessed October 26, 2023.</ref> - Manufacturer of various MEMS sensors.
* [[Analog Devices]]
*   [[Analog Devices]] - Often provides higher-grade IMUs.
* [[Xsens]]
*   [[Xsens]] - Specializes in high-precision motion tracking modules often incorporating IMUs.<ref>Yole Développement. "MEMS & Sensors for Wearables Report." 2023.</ref>


==Notable IMU Models in VR/AR==
== Notable IMU Models in VR/AR ==
*   '''MPU-6050''': A popular low-cost 6-axis IMU (accelerometer + gyroscope) from InvenSense, used in hobbyist projects and early devices like the [[Oculus Rift DK1]].<ref name="MPU6050_Datasheet"> InvenSense Inc. MPU-6000 and MPU-6050 Product Specification Revision 3.4. [https://invensense.tdk.com/wp-content/uploads/2015/02/MPU-6000-Datasheet1.pdf Datasheet Link]</ref>
* '''MPU‑6050''': Low‑cost 6‑axis IMU, used in early [[Oculus Rift]] prototypes.<ref name="MPU6050_Datasheet">*MPU‑6000/6050 Product Specification* Rev 3.4. InvenSense Inc., 2013. https://invensense.tdk.com/wp-content/uploads/2015/02/MPU-6000-Datasheet1.pdf</ref>
*   '''MPU-9250''': An InvenSense 9-axis IMU (adds a magnetometer to the MPU-6xxx series capabilities).<ref name="MPU9250_Datasheet"> InvenSense Inc. MPU-9250 Product Specification Revision 1.1. [https://invensense.tdk.com/wp-content/uploads/2015/02/PS-MPU-9250A-01-v1.1.pdf Datasheet Link]</ref> Used in some dev kits and controllers.
* '''MPU‑9250''': 9‑axis IMU adding a magnetometer.<ref name="MPU9250_Datasheet">*MPU‑9250 Product Specification* Rev 1.1. InvenSense Inc., 2016. https://invensense.tdk.com/wp-content/uploads/2015/02/PS-MPU-9250A-01-v1.1.pdf</ref>
*   '''ICM-42688-P''': A high-performance 6-axis IMU from TDK InvenSense, known for its low noise and stability, used in the Meta [[Quest 2]] headset.<ref name="Quest2_Teardown_iFixit"> iFixit. Meta Quest 2 Teardown. [https://www.ifixit.com/Teardown/Oculus+Quest+2+Teardown/137613 iFixit Teardown]. Accessed October 26, 2023. (Identifies the TDK ICM-42688-P chip)</ref>
* '''ICM‑42688‑P''': High‑performance 6‑axis IMU in the Meta [[Quest 2]].<ref name="Quest2_Teardown_iFixit">iFixit (2021) “Meta Quest 2 Teardown: Into the Metaverse.” https://www.ifixit.com/News/56892/oculus-quest-2-teardown-into-the-metaverse</ref>
*   '''BMI085/BMI270''': Bosch IMUs optimized for VR/AR applications, found in devices like the [[Valve Index]] controllers.<ref>Nield, David. "How VR Headsets Are Getting Better Through Improved Tracking." TechRadar, 2022.</ref>
* '''BMI085/BMI270''': Bosch IMUs used in [[Valve Index]] controllers.<ref>Nield, D. (29 May 2022) “These realityOS trademarks hint at an imminent Apple AR/VR headset launch.” TechRadar. https://www.techradar.com/news/these-realityos-trademarks-hint-at-an-imminent-apple-arvr-headset-launch</ref>
*   '''LSM6DSO/LSM6DSOX''': STMicroelectronics 6-axis IMUs used in various HMDs and AR glasses, including the [[HoloLens 2]].
* '''LSM6DSO(X)''': STMicroelectronics 6‑axis IMUs found in the [[HoloLens 2]].


==References==
== References ==
<references />
<references />


[[Category:Terms]] [[Category:Technical Terms]] [[Category:Tracking]] [[Category:Tracking Technology]] [[Category:Hardware]]
[[Category:Terms]][[Category:Technical terms]][[Category:Tracking]][[Category:Tracking technology]][[Category:Hardware]]

Revision as of 07:40, 3 May 2025

See also: Terms and Technical terms

An Inertial Measurement Unit (IMU) is an electronic sensor device that measures and reports a body's specific force, angular rate, and sometimes orientation using a combination of accelerometers, gyroscopes, and often magnetometers.[1] IMUs are fundamental components in Virtual Reality (VR) and Augmented Reality (AR) systems, tracking the orientation of HMDs and Input Devices like controllers.

Components and Function

A typical IMU integrates multiple sensor types onto a microchip:

An IMU containing all three sensors is called a 9‑axis IMU or a MARG (Magnetic, Angular‑Rate, and Gravity) sensor.[3]

Sensor Fusion

Raw data are noisy or drift; Sensor fusion algorithms such as complementary or Kalman filters combine the signals to yield a stable estimate.[4]

Role in VR/AR

IMUs provide the sub‑millisecond rotational tracking needed for immersion and for avoiding motion sickness.[5]

6DoF Tracking Systems

For full 6DoF tracking, IMU data are fused with camera‑based inside‑out tracking or external lighthouse tracking systems to correct long‑term drift.[6]

Limitations and Correction

  • Sensor drift from gyroscopes
  • Magnetic interference
  • No positional data without external references

Drift is mitigated by visual corrections, complementary filters, and periodic resets such as Zero‑velocity updates (ZUPTs).[7]

IMU Specifications for VR/AR

  • Latency < 2 ms
  • Update rate 500–1000 Hz
  • Gyro drift < 0.01 °/s
  • Low‑noise accelerometers
  • Integrated processing / basic fusion[8]

Future Developments

Trends include lower power draw, on‑sensor machine‑learning, and reduced drift.[9]

Key IMU Manufacturers

Notable IMU Models in VR/AR

  • MPU‑6050: Low‑cost 6‑axis IMU, used in early Oculus Rift prototypes.[12]
  • MPU‑9250: 9‑axis IMU adding a magnetometer.[13]
  • ICM‑42688‑P: High‑performance 6‑axis IMU in the Meta Quest 2.[14]
  • BMI085/BMI270: Bosch IMUs used in Valve Index controllers.[15]
  • LSM6DSO(X): STMicroelectronics 6‑axis IMUs found in the HoloLens 2.

References

  1. ↑ “MEMS Motion Sensors – Inertial Measurement Units.” TDK InvenSense. https://invensense.tdk.com/ (accessed 2 May 2025).
  2. ↑ 2.0 2.1 2.2 Woodman, O. J. (2007) An Introduction to Inertial Navigation. Technical Report UCAM‑CL‑TR‑696, University of Cambridge Computer Laboratory. https://www.cl.cam.ac.uk/techreports/UCAM-CL-TR-696.pdf
  3. ↑ Madgwick, S. O. H.; Harrison, A. J. L.; Vaidyanathan, R. (2011) “Estimation of IMU and MARG Orientation Using a Gradient Descent Algorithm.” Proc. IEEE ICORR, 1‑7. https://vigir.missouri.edu/~gdesouza/Research/Conference_CDs/RehabWeekZürich/icorr/papers/Madgwick_Estimation_of_IMU_and_MARG_orientation_using_a_gradient_descent_algorithm_ICORR2011.pdf
  4. ↑ Mahony, R.; Hamel, T.; Pflimlin, J.‑M. (2008) “Nonlinear Complementary Filters on the Special Orthogonal Group.” IEEE T‑AC 53 (5): 1203‑1218. https://doi.org/10.1109/TAC.2008.923738
  5. ↑ LaValle, S. M. (2016) Virtual Reality – Chapter 9. https://lavalle.pl/vr/vrch9.pdf
  6. ↑ Ercan, A. O.; Erdem, A. T. (2011) “On Sensor Fusion for Head Tracking in Augmented Reality Applications.” Proc. ACC, 1286‑1291. https://doi.org/10.1109/ACC.2011.5991077
  7. ↑ Cadena, C. et al. (2016) “Past, Present, and Future of SLAM.” arXiv:1606.05830.
  8. ↑ 8.0 8.1 “Motion Sensors – Product Portfolio.” TDK InvenSense. https://invensense.tdk.com/products/motion-tracking/ (accessed 2 May 2025).
  9. ↑ Yole Group (2024) Status of the MEMS Industry 2024 (report brochure). https://www.yolegroup.com/product/report/status-of-the-mems-industry-2024/
  10. ↑ “Inertial Measurement Units (IMUs).” Bosch Sensortec. https://www.bosch-sensortec.com/products/motion-sensors/imus/
  11. ↑ “MEMS and Sensors.” STMicroelectronics. https://www.st.com/en/mems-and-sensors.html
  12. ↑ *MPU‑6000/6050 Product Specification* Rev 3.4. InvenSense Inc., 2013. https://invensense.tdk.com/wp-content/uploads/2015/02/MPU-6000-Datasheet1.pdf
  13. ↑ *MPU‑9250 Product Specification* Rev 1.1. InvenSense Inc., 2016. https://invensense.tdk.com/wp-content/uploads/2015/02/PS-MPU-9250A-01-v1.1.pdf
  14. ↑ iFixit (2021) “Meta Quest 2 Teardown: Into the Metaverse.” https://www.ifixit.com/News/56892/oculus-quest-2-teardown-into-the-metaverse
  15. ↑ Nield, D. (29 May 2022) “These realityOS trademarks hint at an imminent Apple AR/VR headset launch.” TechRadar. https://www.techradar.com/news/these-realityos-trademarks-hint-at-an-imminent-apple-arvr-headset-launch
Retrieved from "https://vrarwiki.com/index.php?title=IMU&oldid=34987"