US20100105479A1

US20100105479A1 - Determining orientation in an external reference frame

Info

Publication number: US20100105479A1
Application number: US12/256,747
Authority: US
Inventors: Andrew Wilson; Steven Michael Beeman
Original assignee: Microsoft Corp
Current assignee: Microsoft Technology Licensing LLC
Priority date: 2008-10-23
Filing date: 2008-10-23
Publication date: 2010-04-29
Also published as: US20100103269A1; US8282487B2

Abstract

Orientation in an external reference is determined. An external-frame acceleration for a device is determined, the external-frame acceleration being in an external reference frame relative to the device. An internal-frame acceleration for the device is determined, the internal-frame acceleration being in an internal reference frame relative to the device. An orientation of the device is determined based on a comparison between a direction of the external-frame acceleration and a direction of the internal-frame acceleration.

Description

BACKGROUND

A gyroscope can use angular momentum to assess a relative orientation of a device in a frame of reference that is internal to that device. However, even the most accurate gyroscopes available may accumulate small orientation errors over time.

SUMMARY

This Summary is provided to introduce a selection of concepts in a simplified form that are further described below in the Detailed Description. This Summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to limit the scope of the claimed subject matter. Furthermore, the claimed subject matter is not limited to implementations that solve any or all disadvantages noted in any part of this disclosure.
Determining orientation in an external reference frame is disclosed herein. An external-frame acceleration for a device is determined, the external-frame acceleration being in an external reference frame relative to the device. An internal-frame acceleration for the device is also determined, the internal-frame acceleration being in an internal reference frame relative to the device. An orientation of the device is determined based on a comparison between a direction of the external-frame acceleration and a direction of the internal-frame acceleration.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1A schematically shows an orientation-determining computing system in accordance with an embodiment of the present disclosure.

FIG. 1B schematically shows a position-determining computing system in accordance with another embodiment of the present disclosure.

FIG. 2 shows an exemplary configuration of the orientation-determining computing system of FIG. 1.

FIG. 3 shows a comparison of an external-frame acceleration vector and an internal-frame acceleration vector corresponding to the controller orientation of FIG. 2.

FIG. 4 shows a process flow of an example method of tracking an orientation of a game controller.

DETAILED DESCRIPTION

FIG. 1 shows an orientation-determining computing system 10 including a wand 12, a wand monitor 14 and an orientation inferring subsystem 16. Orientation inferring subsystem 16 is configured to determine an orientation of wand 12 in a frame of reference that is external to the wand 12. In particular, the orientation inferring subsystem 16 may infer a coarse orientation of the wand 12 in the external reference frame by comparing acceleration information of the wand 12 in the external reference frame with acceleration information of the wand 12 in an internal reference frame.
The acceleration information in the external reference frame may be assessed by wand monitor 14. The wand monitor 14 may be configured to observe the wand 12 as the wand 12 moves relative to the wand monitor 14. Such observations may be translated into an external-frame acceleration for the wand. Any suitable technique may be used by the wand monitor 14 for observing the wand 12. As a nonlimiting example, the wand monitor 14 may be configured to visually observe the wand 12 with stereo cameras. In some embodiments, the wand 12 may include a target 18 that facilitates observation by the wand monitor 14.
The acceleration information in the internal reference frame may be assessed by the wand 12. The wand 12 may be configured to sense wand accelerations and report such sensed accelerations to orientation inferring subsystem 16. In some embodiments, the wand may include an acceleration-measuring subsystem 20 for measuring wand accelerations in a frame of reference that is internal to the wand 12.
In addition to determining a coarse orientation of the wand 12 by comparing wand accelerations in internal and external reference frames, the orientation inferring subsystem 16 may update the coarse orientation of the wand 12 based on angular motion information observed by the wand 12 itself. As such, the wand 12 may include an angular-motion measuring subsystem 22 for measuring angular motion of the wand 12 in a frame of reference that is internal to the wand. Even when such an angular-motion measuring subsystem 22 is included, the coarse orientation inferred using internal and external-frame accelerations may be used to limit errors that may accumulate if only the angular-motion measuring subsystem 22 is used.
The wand may be configured to serve a variety of different functions in different embodiments without departing from the scope of this disclosure. As a nonlimiting example, in some embodiments, computing system 10 may be a game system in which wand 12 is a game controller device for controlling various game functions. It is to be understood that the orientation inferring methods described herein may additionally and/or alternatively be applied to an orientation-determining computing system other than a game system, and the wand need not be a game controller in all embodiments.
Furthermore, it is to be understood that the arrangement shown in FIG. 1A is exemplary, and other arrangements are within the scope of this disclosure. As a nonlimiting example, FIG. 1B shows a position-determining computing system 10′ in accordance with another embodiment of the present disclosure. Position-determining computing system 10′ includes a wand 12′, a target monitor 14′ and a position inferring subsystem 16′. Position inferring subsystem 16′ is configured to determine a position of wand 12′ in a frame of reference that is external to the wand 12′. In particular, the position inferring subsystem 16′ may infer a coarse position of the wand 12′ in the external reference frame by comparing orientation information of the wand 12′ in the external reference frame with acceleration information of the wand 12′ in an internal reference frame.
In some embodiments, target 18′ may include one or more LEDs (e.g., infrared LEDs) positioned in a fixed location, such as near a television or any other suitable location. In such embodiments, the wand 12′ may include a target monitor 14′ configured to view the target 18′ and deduce an orientation of the wand based upon a relative position of the target 18′ within the target monitor's field of view. Such information may be used in cooperation with acceleration information measured by an acceleration-measuring subsystem 20′ and/or angular motion information measured by an angular-motion measuring subsystem 22′ to infer a coarse position of the wand as discussed below with reference to inferring coarse orientation.
In yet other embodiments, a wand may include both a target and a target monitor, and/or both a target and a target monitor may be positioned at one or more locations external to the wand. In other words, the arrangements shown in FIGS. 1A and 1B may be at least partially combined, thus enabling direct deduction of both wand position and wand orientation, which may optionally be confirmed/verified with inferred position and inferred orientation, as described herein. Further, it should be understood that the relative positioning of targets, target monitors, wand monitors, and other components described herein may be varied from the specific examples provided herein without departing from the scope of the present disclosure.
FIG. 2 shows an example game system 30 including a controller 32, a controller monitor 34 including stereo cameras 36, and a gaming console 38 including an orientation inferring subsystem 40.
In such a game system 30, orientation inferring subsystem 40 is configured to infer a coarse orientation of controller 32 in an external reference frame relative to controller 32. In particular, the coarse orientation of the controller 32 in a television's, or other display's, reference frame may be inferred. The orientation inferring subsystem 40 infers the coarse orientation of the controller 32 by comparing acceleration information from an external reference frame relative to the controller 32 with acceleration information from an internal reference frame relative to the controller 32.
In the illustrated embodiment, orientation inferring subsystem 40 is configured to determine an external-frame acceleration of controller 32 using time-elapsed position information received from stereo cameras 36. While shown placed near a television, it should be understood that stereo cameras 36, or another wand/target monitor, may be placed in numerous different positions without departing from the scope of this disclosure.
The stereo cameras may observe a target 41 in the form of an infrared light on controller 32. The individual position of the target 41 in each camera's field of view may be cooperatively used to determine a three-dimensional position of the target 41, and thus the controller 32, at various times. Visually-observed initial position information and subsequent position information may be used to calculate the external-frame acceleration of the controller 32 using any suitable technique.
The following technique is a nonlimiting example for using initial position information and subsequent position information to determine an external-frame acceleration of the controller. Taking X₀ to be a current position of controller 32 as observed by controller monitor 34 at a time t₀, and X₋₁ to be a previous position of controller 32 as observed by controller monitor 34 at a previous time t₋₁, an expected position X₀′ for controller 32 at a current time t₀can be calculated according to the following equation,
X ₀′= X ₋₁ + V (t ₀ −t ₋₁).
Here, the velocity V is calculated from prior position information as follows,
$\overline{V} = \frac{(\overline{X_{- 1}} - \overline{X_{- 2}})}{(t_{- 1} - t_{- 2})},$
where X₋₂ is a more previous position of the controller as observed by the controller monitor at a more previous time t₋₂.
If it is determined that the expected position X₀′ is not equal to the current position X₀ , then the difference may be a result of acceleration of controller 32. In such a case, the orientation inferring subsystem 40 determines an external-frame acceleration ā of controller 32 at a current time t₀to be given by the following,
$\overline{a} = \frac{2 (\overline{X_{0}} - \overline{X_{0}^{'}})}{{(t_{0} - t_{- 1})}^{2}} + \overline{g},$
where g is a gravitational acceleration.
Orientation inferring subsystem 40 is configured to determine an internal-frame acceleration of controller 32 from acceleration information received from controller 32. The controller 32 may obtain the internal-frame acceleration in any suitable manner. For example, the controller may include an acceleration-measuring subsystem configured to report acceleration information to the orientation inferring subsystem 40. In some embodiments, the acceleration-measuring subsystem may be a three-axis accelerometer 42 located proximate to the target 41, as schematically shown in FIG. 2.
The orientation inferring subsystem 40 can determine a coarse orientation of controller 32 based on a comparison between a direction of the external-frame acceleration and a direction of the internal-frame acceleration. FIG. 3 shows an example of such a comparison 50 corresponding to the controller movement shown in FIG. 2. Vector 52 represents the direction of the external-frame acceleration and vector 54 represents the direction of the internal-frame acceleration. The misalignment between the external-frame acceleration and the internal-frame acceleration can be resolved to find any difference between the external reference frame and the internal reference frame. Accordingly, an orientation of the controller 32 can be inferred in the external frame of reference.
As a nonlimiting example, if stereo cameras 36 observe controller 32 accelerating due east without changing elevation or moving north/south; and if acceleration-measuring subsystem 20 reports that controller 32 accelerates to the right, without moving up/down or front/back; then orientation inferring subsystem 40 can infer that controller 32 is pointing toward the north. The above is a simplified and somewhat exaggerated scenario. In many usage scenarios, controller 32 will be pointed substantially toward a television or other display, and any relative misalignments between internal and external reference frames will be less severe. Nonetheless, the orientation inferring methods described herein may be used to assess a coarse orientation.
The assessed external-frame acceleration of controller 32 may differ from the actual controller acceleration due to one or more of the following factors: noise and error in the data visually-observed by stereo cameras 36, noise and error in the accelerometer data, and/or misalignment between the internal reference frame and the external reference frame. However, an inferred coarse orientation of controller 32, which is found as described herein, is absolute, rather than relative, and therefore does not accumulate error over time.
In some embodiments, orientation inferring subsystem 40 may be further configured to update the coarse orientation of controller 32 based on angular motion information observed by controller 32. The controller 32 may obtain the angular motion information in any suitable manner. One such suitable manner includes obtaining the angular motion information by means of an angular-motion measuring subsystem 44 configured to report angular motion information to the orientation inferring subsystem 40. In some embodiments, the angular-motion measuring subsystem may include spaced-apart three-axis accelerometers configured to be used in combination to determine the angular motion of controller 32. As shown in FIG. 2, in such embodiments, one three-axis accelerometer 42 may be located at a head end of controller 32 and another three-axis accelerometer 46 may be located at a tail end of controller 32, such that subtracting a head acceleration direction obtained by the head accelerometer 42 from a tail acceleration direction obtained by the tail accelerometer 46 yields an orientation change of controller 32 in the internal reference frame relative to controller 32. In other embodiments, such an angular-motion measuring subsystem 44 may include a three-axis gyroscope 48 which calculates the angular velocity of controller 32, which can then be integrated over time to determine an angular position.
In between frames where a coarse orientation is available (e.g., if target 41 does not move sufficient distance for detection by stereo cameras 36), measurements from the angular-motion measuring subsystem 44 may accumulate error. A long period of very slow motion, as might well happen when drawing, is the worst-case scenario. However, such a situation is the best-case scenario for smoothing and filtering the accelerometer data, because it is expected that a user will attempt to draw smooth lines and curves.
Controller 32 may report acceleration information and/or angular motion information to orientation inferring subsystem 40 by any suitable means. In some embodiments, controller 32 may report acceleration information and/or angular motion information by wirelessly transmitting such information to orientation inferring subsystem 40, as schematically shown in FIG. 2. In other embodiments, controller 32 may be physically connected to orientation inferring subsystem 40.
FIG. 4 shows a process flow diagram of an example method 60 of tracking an orientation of a game controller. Method 60 begins at 62 by inferring a coarse orientation of the game controller. At 64, method 60 includes determining an external-frame acceleration for the game controller, the external-frame acceleration being in an external reference frame relative to the game controller. At 66, method 60 includes determining an internal-frame acceleration for the game controller, the internal-frame acceleration being in an internal reference frame relative to the game controller. At 68, method 60 includes determining an orientation of the game controller based on a comparison between a direction of the external-frame acceleration and a direction of the internal-frame acceleration, as explained above. Upon inferring a coarse orientation of the game controller, method 60 may optionally include, at 70, updating the coarse orientation of the game controller based on angular motion information observed by the game controller.
In some embodiments, an unscented Kalman filter may be used to combine three-dimensional position tracking from stereo cameras, angular velocity information from gyroscopes, and acceleration information from accelerometers into a unified estimate of position and absolute orientation of the device. An unscented Kalman filter may be appropriate because of nonlinearities that may be introduced in the observation part of the process model (i.e., using the orientation to correct accelerometers). An extended Kalman filter may alternatively be used.
The Kalman filter approach combines the information provided from all sensors and allows the introduction of (Gaussian) noise models for each of the sensors. For example, any noise associated with position estimates from the cameras can be incorporated directly into the model. Similarly, the noise of the gyroscopes and accelerometers may be represented by the model. By tuning each of these separately, the system may favor the more reliable sensors without neglecting less reliable sensors.
The Kalman state, state transition, and observation model are described as follows, and the standard Kalman filter equations are used thereafter. At each frame, the state is updated with the state transition model, and predicted sensor values are computed from state estimates given the observation model. After the filter is updated, an updated position and orientation information is “read” from the updated state vector.
The Kalman state {x, {dot over (x)}, {umlaut over (x)}, q, ω} includes information to be represented and carried from frame to frame, and is described as follows:

- x is a 3D position of the device (3-vector);
- {dot over (x)} is a velocity of the device (3-vector);
- {umlaut over (x)} is an acceleration of the device (3-vector);
- q is a device orientation (quaternion); and
- ω is an angular velocity: change in yaw, pitch and roll in the device coordinate frame (3-vector).

Next, a state transition is used to advance the state to the next time step based on process dynamics (velocity, acceleration, etc.). The state transition is described mathematically as follows:
x′=x+{dot over (x)}
{dot over (x)}′={dot over (x)}+{umlaut over (x)}
{umlaut over (x)}′={umlaut over (x)}
q′=q·q(ω)
where:

- q(ω) is a quaternion formed from a change in yaw, pitch, and roll.

Next the sensed values are “observed” from the state, as follows:

- z is a 3D position from a stereo camera system (3-vector);
- gyro are gyroscope values including change in yaw, pitch and roll (3-vector);
- a is accelerometer values (3-vector);
- g is a direction of gravity (3-vector);

where:

- z=x;
- gyro=ω;
- a=({umlaut over (x)}−g)R(q)

where:

- R(q) is a rotation matrix formed from the quaternion q.

The last equation is the focus, where the accelerometer values are predicted by combining the effects of acceleration due to motion of the device, the effect of gravity, and the absolute orientation of the device. Discrepancies in the predicted values are then propagated back to the state by way of the standard Kalman update equations.
It should be understood that the configurations and/or approaches described herein are exemplary in nature, and that these specific embodiments or examples are not to be considered in a limiting sense, because numerous variations are possible. The specific routines or methods described herein may represent one or more of any number of processing strategies. As such, various acts illustrated may be performed in the sequence illustrated, in other sequences, in parallel, or in some cases omitted. Likewise, the order of the above-described processes may be changed.
The subject matter of the present disclosure includes all novel and nonobvious combinations and subcombinations of the various processes, systems and configurations, and other features, functions, acts, and/or properties disclosed herein, as well as any and all equivalents thereof. Furthermore, U.S. Pat. No. 6,982,697 is hereby incorporated herein by reference for all purposes.

Claims

1. A game system, comprising:

a controller;

a controller monitor; and

an orientation inferring subsystem configured to:

determine an external-frame acceleration of the controller from time-elapsed position information received from the controller monitor, the external-frame acceleration being in an external reference frame relative to the controller;

determine an internal-frame acceleration for the device from acceleration information received from the controller, the internal-frame acceleration being in an internal reference frame relative to the controller; and

determine a coarse orientation of the controller based on a comparison between a direction of the external-frame acceleration and a direction of the internal-frame acceleration.

2. The game system of claim 1, where the controller includes an acceleration-measuring subsystem configured to report acceleration information to the orientation inferring subsystem.

3. The game system of claim 1, where the controller includes an angular-motion measuring subsystem configured to report angular motion information to the orientation inferring subsystem.

4. The game system of claim 3, where the angular-motion measuring subsystem includes spaced-apart three-axis accelerometers.

5. The game system of claim 3, where the angular-motion measuring subsystem includes a three-axis gyroscope.

6. The game system of claim 3, where the orientation inferring subsystem is configured to update the coarse orientation based on the angular motion information.

7. The game system of claim 1, where the controller monitor includes stereo cameras.

8. The game system of claim 7, where the controller includes an infrared light and the stereo cameras are configured to view the infrared light.

9. The game system of claim 1, where the orientation inferring subsystem determines the external-frame acceleration as:

\frac{2 (\overline{X_{0}} - \overline{X_{0}^{'}})}{{(t_{0} - t_{- 1})}^{2}} + \overline{g}

where:

X₀ is a current position of the controller as observed by the controller monitor at a time t₀;

g is a gravitational acceleration;

X₀′ is X₋₁ + V(t₀−t₋₁)

where:

X₋₁ is a previous position of the controller as observed by the controller monitor at a previous time t₋₁;

\overline{V} is \frac{(\overline{X_{- 1}} - \overline{X_{- 2}})}{(t_{- 1} - t_{- 2})}

where:

X₋₂ is a more previous position of the controller as observed by the controller monitor at a more previous time t₋₂.

10. The game system of claim 1, where the orientation inferring subsystem uses an unscented Kalman filter to determine a unified estimate of position and an absolute orientation of the controller.

11. A method of tracking an orientation of a game controller, the method comprising:

inferring a coarse orientation of the game controller by:

determining an external-frame acceleration for the game controller, the external-frame acceleration being in an external reference frame relative to the game controller;

determining an internal-frame acceleration for the game controller, the internal-frame acceleration being in an internal reference frame relative to the game controller; and

determining an orientation of the game controller based on a comparison between a direction of the external-frame acceleration and a direction of the internal-frame acceleration; and

updating the coarse orientation of the game controller based on angular motion information observed by the game controller.

12. The method of claim 11, where determining an external-frame acceleration for the game controller includes translating motion information for the game controller that is visually observed by a stereo camera.

13. A method of inferring device orientation in an external reference frame, the method comprising:

determining an external-frame acceleration for the device, the external-frame acceleration being in an external reference frame relative to the device;

determining an internal-frame acceleration for the device, the internal-frame acceleration being in an internal reference frame relative to the device;

determining an orientation of the device based on a comparison between a direction of the external-frame acceleration and a direction of the internal-frame acceleration.

14. The method of claim 13, where determining an external-frame acceleration for the device includes translating visually-observed motion of the device.

15. The method of claim 14, where a stereo camera is used to visually-observe motion of the device.

16. The method of claim 13, where determining the internal-frame acceleration for the device includes receiving internal-frame acceleration information observed by the device.

17. The method of claim 13, further comprising updating the orientation of the device based on angular motion information observed by the device.

18. The method of claim 13, where determining an external-frame acceleration for the device includes receiving initial position information for the device, the initial position information being in the external reference frame relative to the device; and receiving subsequent position information for the device, the subsequent position information being in the external reference frame relative to the device.

19. The method of claim 13, where determining the external-frame acceleration for the device includes calculating:

\frac{2 (\overline{X_{0}} - \overline{X_{0}^{'}})}{{(t_{0} - t_{- 1})}^{2}} + \overline{g}

where:

X₀ is a current position of the device in the external reference frame at a time t₀;

g is a gravitational acceleration;

X₀′ is X₋₁ + V(t₀−t₋₁)

where:

X₋₁ is a previous position of the device in the external reference frame at a previous time t₋₁;

\overline{V} is \frac{(\overline{X_{- 1}} - \overline{X_{- 2}})}{(t_{- 1} - t_{- 2})}

where:

X₋ 2 is a more previous position of the device in the external reference frame at a more previous time t₋₂.

20. The method of claim 13, further comprising using an unscented Kalman filter to determine a unified estimate of position and an absolute orientation of the device.