Method And System For Estimating Relative Angle Between Headings

Abstract

The present invention relates to a method for estimating a relative angle (θ) between heading (X.sub.10) of a person (10) and heading (X.sub.12) of a mobile device (12) carried by the person, which method comprises: determining a first estimate of the relative angle using a multi-axis accelerometer (18) of the mobile device; determining a second estimate of the relative angle using a multi-axis gyroscope (20) of the mobile device; and combining the first estimate and the second estimate to a combined estimate of the relative angle (θ) between the person's heading and the mobile device's heading. The present invention also relates to a system for estimating a relative angle (θ) between heading (X.sub.10) of a person (10) and heading (X.sub.12) of a mobile device (12) carried by the person.

Claims

1.-10. (canceled)

11. A method for estimating a relative angle between heading of a person and heading of a mobile device carried by the person, which method comprises: determining a first estimate of the relative angle using a multi-axis accelerometer of the mobile device; determining a second estimate of the relative angle using a multi-axis gyroscope of the mobile device; and combining the first estimate and the second estimate to a combined estimate of the relative angle between the person's heading and the mobile device's heading.

12. A method according to claim 11, wherein determining a first estimate of the relative angle using a multi-axis accelerometer of the mobile device comprises: estimating horizontal acceleration of the mobile device in first and second orthogonal directions when the person is walking or running based on samples from the multi-axis accelerometer; detecting beginning and end of a step of the person from the estimated horizontal accelerations; integrating the estimated horizontal acceleration in the first direction and in the second direction from the detected beginning to the detected end of the step, to provide corresponding velocities in the first and second directions; calculating the norm of the integrated accelerations; finding maximum of the calculated norm of the integrated accelerations; and calculating the first estimate of the relative angle based on the proportion between the velocity in the first direction and the velocity in the second direction for a sample corresponding to said maximum.

13. A method according to claim 12, wherein detecting beginning and end of a step of the person comprises: calculating the norm of the estimated horizontal accelerations; identifying a peak of the calculated norm of the estimated horizontal accelerations, which peak is over a predetermined threshold; selecting the closest local minimum of the calculated norm before the identified peak as beginning of the step; and selecting the closest local minimum of the calculated norm after the identified peak as end of the step.

14. A method according to claim 11, wherein determining a second estimate of the relative angle using a multi-axis gyroscope of the mobile device comprises: estimating an axis that the mobile device is rotating about with respect to pendulous motion; estimating the heading of the person as a direction which is perpendicular to the estimated axis in a ground plane; and obtaining the second estimate of the relative angle by comparing the estimated heading of the person to a projection of the mobile device's heading on the ground plane.

15. A method according to claim 11, wherein the first estimate and the second estimate are combined using Kalman filtering.

16. A method according to claim 11, which method further comprises: detecting absolute heading of the mobile device by using a compass; and estimating absolute heading of the person based on the detected absolute heading of the mobile device and said combined estimate of the relative angle.

17. A method according to claim 11, which method further comprises: providing the combined estimate of the relative angle as an output signal to an indoor positioning system or a pedestrian dead reckoning system.

18. A system for estimating a relative angle between heading of a person and heading of a mobile device carried by the person, which system comprises: the mobile device, wherein the mobile device comprises a multi-axis accelerometer and a multi-axis gyroscope; and a processor capable of communicating with the multi-axis accelerometer and the multi-axis gyroscope, wherein the processor is configured to: determine a first estimate of the relative angle based on data from the multi-axis accelerometer; determine a second estimate of the relative angle based on data from the multi-axis gyroscope of the mobile device; and combine the first estimate and the second estimate to a combined estimate of the relative angle between the person's heading and the mobile device's heading.

19. A system according to claim 18, wherein said processor is included in the mobile device.

20. A system according to claim 18, wherein said processor is included in a remote device of the system.

Description

[0017] These and other aspects of the present invention will now be described in more detail, with reference to the appended drawings showing one or more embodiments of the invention.

[0018] FIG. 1 is top view of a person and a mobile device.

[0019] FIG. 2 is a schematic front view of a mobile device according to an embodiment of the present invention.

[0020] FIG. 3 is a perspective view of the mobile device of FIG. 2.

[0021] FIG. 4 is a flow chart of a method according to an embodiment of the present invention.

[0022] FIG. 5 is a perspective view similar to FIG. 4.

[0023] FIG. 6 is a diagram of norm acc over time.

[0024] FIG. 7 is a diagram of velocities over time.

[0025] FIGS. 8a-8c are schematic side view of a person's gait.

[0026] FIG. 9 is a top view of the device carried by the person of FIG. 8c.

[0027] FIG. 10 shows a system comprising a remote device according to an embodiment of the present invention.

[0028] With reference to FIGS. 1-4, the present invention relates to a method for estimating the relative angle θ between heading X.sub.10 of a person 10 and heading X.sub.12 of a mobile device 12 carried by the person 10. The method is computer-implemented. The heading X.sub.10 of the person 10 is typically the same as the person's course when the person is walking or running. The course of the mobile device 12 is the same as the course of the person 10, but the heading X.sub.12 of the mobile device may indeed be different than its course depending on how the mobile device 12 is carried by the person 10.

[0029] The mobile device 12 (FIG. 2) may be a mobile communication device, such as a mobile phone or smart phone. The mobile device 12 has a housing 14 and a screen 16. The housing 14 may have the general shape of a rectangular box. The mobile device 12 further comprises a multi-axis accelerometer 18 and a multi-axis gyroscope 20. The multi-axis accelerometer 18 may be a three-axis accelerometer, and the multi-axis gyroscope 20 may be a three-axis gyroscope. The accelerometer 18 and the gyroscope 20 are fixedly mounted in the housing 14 of the mobile device 12. The accelerometer 18 and the gyroscope 20 are connected to a processor 22 of the mobile device 12. The processor 22 is generally configured to perform various method steps of the present invention. The accelerometer 18 is adapted to detect acceleration in three orthogonal axes in the accelerometer's frame of reference, namely x-axis acceleration, y-axis acceleration and z-axis acceleration, as shown in FIG. 3. The gyroscope is adapted to detect rotation about the same axes, namely x-axis rotation (pitch), y-axis rotation (roll) and z-axis rotation (yaw), as also shown in FIG. 3. The mobile device 12 may optionally further comprise a compass 24, for example an electronic compass (e-compass). The compass 24 is also connected to the processor 22.

[0030] In the method (FIG. 4), a first estimate of the relative angle θ is determined (step S1) using the multi-axis accelerometer 18, and a second estimate of the relative angle θ is determined (step S2) using the multi-axis gyroscope 20. The first and second estimates are then combined (step S3) to a combined estimate of the relative angle θ between the heading X.sub.10 of the person 10 and the heading X.sub.12 of the mobile device 12. The first and second estimates may for example be combined using Kalman filtering. By combining the first estimate based on input from the accelerometer 18 and the second estimate based on input from the gyroscope 20, a high accuracy estimation of the relative angle θ may be obtained.

[0031] The method may optionally detect (step S4) the absolute heading of the mobile device 12, by using the compass 24, and estimate the absolute heading of the person 10 based on the detected absolute heading of the mobile device 12 and the combined estimate of the relative angle θ.

[0032] The combined estimate of the relative angle θ and/or the absolute heading of the person may beneficially be provided as an output signal to an indoor positioning system or a pedestrian dead reckoning system.

[0033] Accelerometer Model

[0034] Step S2 of determining the first estimate of the relative angle θ using the multi-axis accelerometer 18 according to an embodiment of the invention will now be described with reference to FIGS. 5-7.

[0035] First, horizontal acceleration of the mobile device 12 in first and second orthogonal directions x′, y′ is estimated based on samples from the multi-axis accelerometer 18 when the person 10 is walking or running. If the mobile device 12 is flush with a horizontal plane 26 as shown in FIG. 5, the first direction and second directions x′, y′ may correspond to the x-axis and y-axis of the accelerometer 18, respectively. However, is the mobile device 12 is tilted with respect to the horizontal plane 26, the acceleration samples from the accelerometer 18 should be transformed from the accelerometer's frame of reference to the Earth or World coordinate frame, which includes the horizontal plane. In other words, the acceleration samples may be “tilt-adjusted”. ‘Horizontal acceleration(s)’ herein typically refers to accelerations in or parallel to a ground plane at the person's location.

[0036] Then, steps of the person 10 are detected by studying the estimated horizontal accelerations of the mobile device 12. Specifically, the norm of the estimated horizontal acceleration in the first direction x′ and the estimated horizontal acceleration in the first direction y′ is calculated:

norm.sub.acc=√{square root over (acc.sub.x′.sup.2+acc.sub.y′.sup.2)}

[0037] An example of the calculated norm over time (samples) is shown in FIG. 6. Thereafter, a peak of the calculated norm of the estimated horizontal accelerations is identified, which peak is over a predetermined threshold. The method then finds where the norm's derivative is 0 before and after the identified peak, wherein the closest local minimum of the calculated norm before the identified peak is selected as the beginning B of the step, and the closest local minimum of the calculated norm after the identified peak is selected as the end E of the step. Several steps of the person are indicated in FIG. 6.

[0038] When beginning B and end E of a step of the person 10 have been detected, the estimated horizontal accelerations in both the first direction x′ and in the second direction y′ are integrated over the complete step. This gives the change in velocity in both the first direction x′ and the second direction y′, see FIG. 7. Then the norm of the integrated accelerations is calculated. This norm has a maximum M, and the sample (time) corresponding to the maximum is used to calculated the first estimate of the relative angle θ. Specifically, the first estimate of the relative angle θ is calculated based on the proportion between the velocity in the first direction x′ and the velocity in the second direction y′ for said sample. For example, if the velocity v.sub.x′ in the first direction x′ is equal to the velocity v.sub.y′ in the second direction y′, then the first estimate of the relative angle θ is 45 deg. This example is illustrated in FIG. 5. Mathematically, the first estimate of the relative angle θ may be calculated as:

[00001]$\arctan \left(\frac{v_{y^{\prime }}}{v_{x^{\prime }}}\right).$

[0039] Pendulum Model

[0040] Step S3 of determining the second estimate of the relative angle using the multi-axis gyroscope 20 according to an embodiment of the invention will now be described with reference to FIGS. 8a-8c and 9.

[0041] The pendulum model is based on the assumption that the mobile device 12 is borne on the leg, hip or other locations (of the person 10) subjected to pendulous motion, and it makes use of the changes in tilt angles that the leg or hip or other location and the mobile device are subjected to. R in FIGS. 8a-8c denotes the right leg of the person 10, and L denotes the left leg.

[0042] Specifically, step S3 may comprise estimating an axis P that the mobile device 10 is rotating about with respect to pendulous motion when the person 10 is walking or running. ω is the rotation around axis P. This axis of rotation P seen in FIGS. 8a-8c is estimated by comparing the orientation of the mobile device 12 to a reference orientation, which for example can be the moving average of the orientation during walk. An estimate of the axis P can be calculated by comparing the latest estimate of the orientation to the previous estimate. These estimates of P can then be low-pass filtered to reduce the effects of measurement noise.

[0043] Given an accurate estimate of the axis P, an estimate of the forward direction X.sub.10 of the person's walk (gait) is given as the axis/direction which is perpendicular to P in the horizontal plane. The forward direction X.sub.10 is then compared to a projection of the mobile device's x-axis X.sub.12 on the horizontal plane to get the second estimate of the relative angle θ, as seen in FIG. 9.

[0044] The mobile device 12 may constitute a system for estimating the relative angle θ between its heading X.sub.12 and the heading X.sub.10 of the person 10 carrying it. Alternatively, such a system may comprise a remote device 28, i.e. not part of the mobile device 12, which remote device 28 in turn comprises a processor 30 configured to perform various method steps of the present invention. Here, less processing power in the mobile device 12 is needed. The alternative system is schematically illustrated in FIG. 10. The processor 28 of the remote device may for example be configured to determine the first and second estimates of the relative angle based on data from the multi-axis accelerometer and the multi-axis gyroscope, respectively, and combine the first estimate and the second estimate to a combined estimate of the relative angle θ. The remote device may for example be a computer or a server capable of communicating with the mobile device 12, for example over a wireless connection.

[0045] The person skilled in the art realizes that the present invention by no means is limited to the preferred embodiments described above. On the contrary, many modifications and variations are possible within the scope of the appended claims.