WO2010082157A1

WO2010082157A1 - Method for determining the rotation axis of a joint and device for monitoring the movements of at least one body part

Info

Publication number: WO2010082157A1
Application number: PCT/IB2010/050112
Authority: WO
Inventors: Victor M. G. Van Acht; Nicolaas Lambert
Original assignee: Koninklijke Philips Electronics NV
Current assignee: Koninklijke Philips NV
Priority date: 2009-01-16
Filing date: 2010-01-12
Publication date: 2010-07-22
Anticipated expiration: 2011-07-16

Abstract

A method for determining the rotation axis (R) of a joint is provided. The method comprises the steps: attaching a position and orientation indicator (2, 4) to a body part (3, 5) adjacent to a joint; moving the joint and detecting resulting motions of the position and orientation indicator (2, 4) with respect to a reference system; determining instantaneous orientations (WQs) of the position and orientation indicator (2, 4) at different points in time from the detected motions of the position and orientation indicator; determining an average orientation (|_^βs J) of the position and orientation indicator from the detected motions of the position and orientation indicator; and determining a rotation axis (R) of the joint based on the instantaneous orientations and the average orientation.

Description

METHOD FOR DETERMINING THE ROTATION AXIS OF A JOINT AND DEVICE FOR MONITORING THE MOVEMENTS OF AT LEAST ONE BODY PART

FIELD OF INVENTION

The present invention relates to a method for determining the rotation axis of a joint and to a device for monitoring the movements of at least one body part. BACKGROUND OF THE INVENTION

Conventionally, physiotherapy is very labor intensive, since a physiotherapist can treat only one person at a time. As a result of this, conventional physiotherapy is rather expensive which results in that only a limited number of treatments is reimbursed to the patient in order to keep costs for the healthcare system low. The outcome for the patient would be improved if he/she could do more or longer physiotherapy exercises, but the costs for the physiotherapist which has to be present all the time prevents prescription of more and/or longer physiotherapy exercises.

In view of this, tools for therapists are being developed which enable the patients to execute physiotherapy exercises with less guidance from the therapist. Recently, it has been proposed to use position and orientation indicators for guiding physiotherapy patients on how to do their exercises in the correct way. According to this technique, the patient wears one or more position and orientation indicators for measuring the patient's movements. A computer evaluates the movements and can provide real-time feedback to the patient to motivate the patient to perform the exercises in the correct way. Further, the computer can generate progress reports for the therapist such that the therapist can guide for example 5 or 10 patients simultaneously working with such a type of system.

In known devices for monitoring the movements of at least one body part, calibration procedures are required to determine the alignment of one or more position and orientation indicators with respect to the body parts to which they are attached. According to a system known to the applicant, the position and orientation indicators are formed by motion sensors attached to different body parts and transmitting position and orientation information to a control unit. It is known to use a specific calibration procedure in which the patient is asked to take a predefined reference posture to acquire information about the alignment of the motion sensor(s) with respect to the body part(s). This procedure is executed before the patient starts exercising. When the patient is in the reference posture (e.g. standing up right with the arms strictly down and the palms of the hands towards the legs) the motion sensor reading is stored as a reference reading. During normal operation of the system (i.e. when the patient performs the exercises), the current reading of the motion sensor can then be corrected using the reference reading (e.g. by "subtracting" the reference reading) in order to calculate the current orientation of the limbs from the orientation of respective motion sensors. However, such an alignment procedure as described above is cumbersome and thus user-unfriendly. Further, if the patient does not accurately assume the reference posture during the calibration procedure (and this is not detected), all calculations based on the accuracy of the alignment calibration fail.

US 2008 0285805 Al discloses a system for capturing motion of a moving object via a plurality of motion sensor modules placed on various body segments. The sensor modules capture both 3D position and 3D orientation data relating to their respective body segments. According to the disclosed system, first a sensor-to-segment alignment has to be performed before the system can be used for capturing the motion of a moving object. Thus, there is a demand to dispense with the explicit calibration procedure.

On the other hand, there is a demand to determine the rotation axis of a joint (such as the hip, the shoulder, the knee, the elbow, and the like) since, when the rotation axis is known, it can e.g. be used to evaluate whether this rotation axis is in the desired direction. This will be exemplarily explained for a hip joint with reference to Fig. 1. In Fig. Ia) lifting a leg forward is shown. In this case, the rotation axis R is perpendicular to the sagittal plane SP. In Fig. Ib) twisting of the leg is shown, i.e. the rotation axis R is perpendicular to the transverse plane TP of the rotating leg. In Fig. Ic) lifting the leg sideward is shown, i.e. the rotation axis R is perpendicular to the coronal plane CP. In such a setup it can be checked whether the limb stays within a desired, pre-defined motion plane or deviates out of such a motion plane at the extremes of the movements.

SUMMARY OF THE INVENTION

It is an object of the present invention to provide information about the actual movements of a body part without requiring an explicit calibration procedure for determining the alignment of a position and orientation indicator with respect to a body part.

This object is solved by a method for determining the rotation axis of a joint according to claim 1. The method comprises the steps: attaching a position and orientation indicator to a body part adjacent to a joint; moving the joint and detecting resulting motions of the position and orientation indicator with respect to a reference system; determining instantaneous orientations of the position and orientation indicator at different points in time from the detected motions of the position and orientation indicator; determining an average orientation of the position and orientation indicator from the detected motions of the position and orientation indicator; and determining a rotation axis of the joint based on the instantaneous orientations and the average orientation. In this way, the actual rotation axis of a joint, i.e. the axis around which a joint is currently rotated, can be reliably determined without requiring an initial position and orientation indicator -to-body part alignment routine. Thus, the position and orientation indicator is attached to the body part without necessitating a specific calibration of its alignment. The reference system with respect to which the movements of the position and orientation indicator are determined can e.g. be a further position and orientation indicator attached to a body part or e.g. a "world-fixed" reference position, e.g. a specific position inside a room or the like. Since the actual orientation of the rotation axis is determined based on the instantaneous orientations and the average orientation of the position and orientation indicator, no calibration procedure is necessary. Preferably, the method comprises the step of calculating the differences between the instantaneous orientations and the average orientation, and the rotation axis of the joint is determined based on these differences. In this case, the rotation axis is determined in a particular convenient way.

According to one aspect, the reference system is a global fixed system. In this case, only one position and orientation indicator has to be attached to the body and the system can be realized in a particularly cost-efficient way.

According to another aspect, the reference system is a second position and orientation indicator attached to a body part on the opposite side of the joint. In this case, the rotation axis of the joint can be reliably determined even if the orientation and the position of the joint are not constant over time (which is a situation during physiotherapy exercises when e.g. the elbow and knee joints are moved while simultaneously moving the shoulder joint respectively the hip joint).

Preferably, the position and orientation indicator is attached to the body part such that rotation about a first axis of a local coordinate frame of the position and orientation indicator with respect to the body part is prevented. In this case, the determined orientation of the rotation axis can conveniently be analyzed to determine components of the rotation axis along specific axes of a local coordinate frame of the body part or of another suitable coordinate frame. If the component of the rotation axis along the first axis of the local coordinate frame of the position and orientation indicator is determined from the differences between the instantaneous orientations of the position and orientation indicator and the average orientation of the position and orientation indicator, for example a twist of a body part during physiotherapy exercises can reliably be determined (for instance twisting of a leg during hip joint exercises or twisting of an arm during shoulder exercises) and this twist can conveniently be expressed in a local coordinate frame.

If the orientation of the rotation axis excluding the component of the rotation axis along the first axis of the local coordinate frame of the position and orientation indicator is determined, the rotation axis of the joint without the twisting of the body part (i.e. with the twisting "subtracted") can be conveniently determined. The object is further solved by a device for monitoring the movements of at least one body part according to claim 8. The device comprises at least one position and orientation indicator adapted to be attached to a body part which is adjacent to a joint and a control unit adapted for detecting motions of the position and orientation indicator with respect to a reference system. The control unit is adapted to: determine instantaneous orientations of the position and orientation indicator at different points in time from the detected motions of the position and orientation indicator; determine an average orientation of the position and orientation indicator from the detected motions of the position and orientation indicator; and determine a rotation axis of the joint based on the instantaneous orientations and the average orientation. Thus, the rotation axis of the joint is reliably determined without requiring explicit calibration of the alignment of the position and orientation indicator with respect to the body part.

Preferably, the device comprises an attachment component for attaching the at least one position and orientation indicator to the body part and the attachment component is structured such that rotation about a first axis of a local coordinate frame of the position and orientation indicator with respect to the body part is prevented. In this case, the determined orientation of the rotation axis can conveniently be analyzed to determine components of the rotation axis along specific axes of a local coordinate frame of the body part or of another suitable coordinate frame.

If the device comprises at least one further position and orientation indicator adapted to be attached to a further body part on the opposite side of the at least one joint; the rotation axis of the joint can be reliably determined even if the position and the orientation of the joint are not fixed over time.

Preferably, the device is a physiotherapy monitoring device. BRIEF DESCRIPTION OF THE DRAWINGS

Further features and advantages of the present invention will arise from the detailed description of embodiments with reference to the enclosed drawings.

Fig. Ia to Ic schematically illustrate rotations of a hip joint around different axes. Fig. 2 schematically shows a human body comprising several body parts provided with position and orientation indicators and associated local coordinate systems.

Fig. 3 schematically illustrates the orientation of a local position and orientation indicator coordinate system.

Fig. 4a to 4c schematically illustrate different degrees of freedom in orientation between local position and orientation indicator coordinate systems and body parts.

Fig. 5 shows an example of a plot used for determining the rotation axis of a joint.

Fig. 6 schematically shows an example of combined lifting and twisting of a leg.

DETAILED DESCRIPTION OF EMBODIMENTS

An embodiment of the present invention will now be described with reference to Figs. 2 to 6. In the embodiment, the method for determining the rotation axis of a joint will be exemplarily explained with respective position and orientation indicators attached to an upper leg and a lower leg of a patient. More clearly, the method will be explained with respect to at least one position and orientation indicator attached to the body part which is a leg in the example. Further, the device for monitoring the movements of at least one body part according to the embodiment is formed by a physiotherapy monitoring device. The device according to the embodiment is specifically adapted for monitoring movements of the leg of a patient, in particular for monitoring hip and knee exercises for physiotherapy patients. However, it should be noted that the invention is not limited to this type of application. For example, movements of other human or animal body parts can be monitored such as e.g. arms, elbows, shoulders, ankles, and the like. In particular, physiotherapeutic exercises can be monitored and/or guided. In the embodiment which will be described in the following, the position and orientation indicators are formed by motion sensors attached to respective body parts and transmitting position and orientation information to the control unit 100. However, it should be noted that the invention is not limited to this and can e.g. also be applied to other types of position and orientation indicators. For example, according to a system known to the applicant, the position and orientation indicators can also be formed by reference elements attached to different body parts, wherein position and orientation of the reference elements are detected by means of a suitable camera. One such system known to the applicant uses reflective balls attached to body parts, illuminating these reflective balls with infrared light, and detecting the position/orientation of the balls by detecting the reflected light with a suitable infrared camera. The device according to the embodiment comprises a control unit 100 (see Fig. 2) which is connected (e.g. through wires or wireless) to the position and orientation indicators which will be described later on and is adapted such that the steps which will be described in following are performed by the device. Fig. 2 schematically shows a human body 1 comprising a plurality of body parts. In the example which will be described in the following, a first position and orientation indicator 2 is attached to an upper leg as a first body part 3 and a second position and orientation indicator 4 is attached to a lower leg as a second body part 5. In the example which will be described, the steps for determining the rotation axis of the joint and further steps are implemented in the control unit 100 which is schematically depicted. A skilled person will understand that rotation axes of other joints can be determined in analogous or similar manner. The first position and orientation indicator 2 and the second position and orientation indicator 4 can e.g. be attached to the first body part 3 and the second body part 5 by means of elastic textile straps comprising small pouches in which the respective position and orientation indicators can be placed. In the example which will be described in the following one position and orientation indicator (first position and orientation indicator 2) is attached "above" the knee (i.e. attached to the upper leg) while another position and orientation indicator (second position and orientation indicator 4) is attached "below" the knee (i.e. attached to the lower leg). In the example shown in Fig. 2, a (right-handed) local orthogonal coordinate frame is assigned to the first body part 3 (the upper leg). This local orthogonal coordinate frame comprises the axes ^ULx, ^ULy, and ^ULz. ^ULz is chosen to point forward in the direction in which the face of the patient wearing the position and orientation indicators is directed (in a normal posture). ^ULx is chosen to point towards the foot, and ^ULy is chosen such that ^ULx, ^ULy, and ^ULz form a right-handed orthogonal coordinate frame. Similarly, a local orthogonal coordinate frame is assigned to the second body part 5 (the lower leg) comprising the axes ^LLx, ^LLy, and ^LLz. ^LLz is chosen to point forward, ^LLx is chosen to point to the foot, and ^LLy is chosen such that a right-handed local orthogonal coordinate frame is formed.

In a similar manner, local coordinate frames are assigned to the first position and orientation indicator 2 and to the second position and orientation indicator 4. The local coordinate frame assigned to the first position and orientation indicator 2 comprises the axes x, y, z. The local coordinate frame assigned to the second position and orientation indicator 4 comprises the axes ^LSx, ^LSy, and ^LSz. The assignment of the local coordinate frames to the first and second position and orientation indicators 2 and 4 is schematically shown in Fig. 3 where x corresponds to ^usx or ^LSx, y to ^usy or ^LSy, and z corresponds to ^usz or ^LSz, respectively. Rx is used to indicate a rotation about the axis x, Ry to indicate a rotation about y, and Rz to indicate a rotation about z.

In a perfect alignment of the first position and orientation indicator 2 to the first body part 3, the direction of ^usx corresponds to ^ULx, the direction of ^usy corresponds to ^ULy, and the direction of ^usz corresponds to ^ULz. Similarly, for a perfect alignment of the second position and orientation indicator 4 to the second body part 5 the direction of

T C T T T C T T T C x corresponds to x, the direction of y corresponds to y, and the direction of z corresponds to ^LLz. Such a perfect alignment situation is schematically depicted in Fig. 2.

Further, in Fig. 2 a local coordinate frame is assigned to the torso 6 with the axis ^τz pointing forward, the axis ^τx pointing towards the head, and ^τy being chosen such that a right-handed orthogonal local coordinate frame is formed. It should be noted that local coordinate frames can be assigned to other body parts in a corresponding way. Further, position and orientation indicators could additionally or alternatively be attached to other body parts. It has been described that, in a perfect alignment situation, the coordinate frames of the first position and orientation indicator 2 ( x, y, z) and the first body part 3 (^ULx, ^ULy, ^ULz) would be equal and the coordinate frames of the second position and orientation indicator 4 ( x, y, z) and the second body part 5 ( x, y, z) would be equal. However, in a real situation in which the position and orientation indicators are attached to respective body parts (e.g. by flexible textile straps), this perfect situation will normally not be present. At least a certain misalignment between the coordinate frames of the first position and orientation indicator 2 and the first body part 3 will be present, i.e. the directions of the corresponding axes will differ. Similarly, a certain misalignment between the coordinate frames of the second position and orientation indicator 4 and the second body part 5 will be present. In the following it will be described how the rotation axis of a joint can be reliably determined even if the position and orientation indicators are not aligned with the body parts to which they are attached. Thus, for the method which will be described in the following, the position and orientation indicators 2, 4 can be attached to the body parts 3, 5 and the rotation axis of the joint can be reliably determined without requiring an explicit calibration procedure before.

In the example given below, determination of the rotation axis of a hip joint will be exemplarily explained based on motions of the leg (e.g. as shown in Fig. 1). First, it will be described in general how the rotation axis is determined based on the position and orientation indicator readings of either the first position and orientation indicator 2 or the second position and orientation indicator 4. It should be noted that the alignment of the position and orientation indicator with respect to the body part is not calibrated before the rotation axis is determined, i.e. the rotation axis is determined using a position and orientation indicator which has not been aligned with the body part to which it is attached.

In the following description, all orientations will be expressed as quaternions. However, the algorithms could also be expressed using another representation of 3 -dimensional orientations such as Euler angles or rotation matrices, albeit with a little more work. First, it will be described how the rotation axis of the joint (the hip joint in the example) can be found. The orientation of the position and orientation indicator (e.g. either the first position and orientation indicator 2 or the second position and orientation indicator 4) expressed in a world- fixed reference coordinate frame (a global coordinate frame which is used as a reference system in the example) is given by the quaternion ^wQs (S for sensor as an example for a position and orientation indicator, W for world), wherein a quaternion ^AQ_B expresses a rotation (or coordinate transformation) from frame B to frame A.

When, during movement of the position and orientation indicator relative to the reference system (which is a motion of the hip joint in the present example), the instantaneous values of ^wQs are passed through a low pass filter with a cut-off frequency of for example 0.3Hz, the average difference in orientation between the position and orientation indicator and the reference system denoted by \_^WQ_S J is obtained.

The difference between the instantaneous value of ^wQs and its average value i Q_s J can be calculated. This difference will be called ^s Q\_s \ ^and is given by the following equation:

where ® denotes the quaternion product and Q* denotes the quaternion conjugate.

The difference ^SQ\_S \ can be interpreted as the instantaneous difference between, on one hand, the current orientation of the position and orientation indicator with respect to the reference system and, on the other hand, the average orientation of the position and orientation indicator with respect to the reference system. And obviously, the difference between the current orientation and the average orientation is caused by movement of the joint around the joint axis in the present example (the current hip joint axis when the only the hip joint has been moved as supposed). Equation (1) expresses the difference in the position and orientation indicator coordinate system, but it can just as well be expressed in the world coordinate system by:

^WQ_LWF^WQS

(2)

When the joint moves, ^SQ\_S \ ^and ^WQ\ _w \ take values on both sides of a line through the origin that expresses exactly the rotation axis of the joint, expressed in the coordinate frame that is fixed to the position and orientation indicator, respectively to the world.

In order to obtain the direction of this rotation axis, the w-component is omitted and the x, y and z component of different samples of the 4-dimensional quaternion

w _\ ^are plotted in a 3 dimensional graph. Then, a plot like the dots in Fig. 5 is obtained. As can be seen in Fig. 5, the dots lie on a line (indicated by a dashed line in Fig. 5). In this representation, the x-, y- and z-components are along the axes.

Now, the next step is to perform a principal component analysis on the x, y and z components of data points of

or ^wQ _w Λo find the main rotation axis (that is, the dashed line in Fig. 5). This is now the rotation axis of the joint. In the example of Fig. 5, the rotation axis would be (x, y, z) = (2, 5, 3). It should be noted that Fig. 5 exemplarily shows either ^SQ\_S \ ^{or W}Q\ w \ ^as dots.

It should be noted that this method does not only apply to find the rotation axis of the hip joint, obviously it can be used to find the rotation axis of any part of the body (appropriately provided with a position and orientation indicator).

Thus far, it has been explained how the rotation of a joint can be determined. The rotation axis could either be expressed in the world coordinate frame (using a principal component analysis of ^WQ\ _W \ ) ^{or m} the position and orientation indicator coordinate frame (using

However, for therapists and patients it is much more intuitive to express the rotation axis of the joint in a combination of the world coordinate system and the position and orientation indicator coordinate system. Namely: For a twist of the leg (rotating knee and foot inward and outward, shown in Fig. Ib) it is natural to express the rotation axis in the local coordinate frame, whereas for lifting the leg (rotating the leg forward or sideway, shown in Fig. Ia and Fig. Ic respectively) it is more natural to express the rotation axis in the world- fixed coordinate frame or in a pelvis-fixed coordinate frame. This becomes most clear for a combined twisting and lifting of the leg: When standing, it is natural to define the twist of the leg along a vertical axis, whereas when the leg is lifted 90 degrees it is natural to define the twist of the leg along a horizontal axis. This is schematically shown in Fig. 6.

Due to this, the result obtained above with respect to the rotation axis of the joint cannot be directly applied to give feedback to a patient or physiotherapist on how to improve the patient's exercising. For this reason, a slight variation of the algorithm from above is made in the following. In the following, the assumption will be made that the position and orientation indicator (e.g. either position and orientation indicator 2 or position and orientation indicator 4) is fixed to the body part such the local x-axis of the position and orientation indicator is along the direction of the body part. This implicitly also assumes that the position and orientation indicator cannot rotate around its local z-axis. Such a fixation can easily be achieved by means of an elastic strap as mentioned above, as will be explained in the following.

It has been described above that the first and second position and orientation indicators 2, 4 can e.g. be attached to the first body part 3 and the second body part 5, respectively, by means of flexible elastic straps comprising pouches for accommodating the respective position and orientation indicators. In this case, at least one rotational degree of freedom of the alignment between the position and orientation indicator and the corresponding body part can be made fixed due to specific attachment. This will be described in more detail with reference to Figs. 4a to 4c for the attachment of the first position and orientation indicator 2 to the upper leg and of the second position and orientation indicator 4 to the lower leg. In the example shown in Figs. 4a to 4c, the first position and orientation indicator 2 is attached to the first body part 3 by means of a strap 7 and the second position and orientation indicator 4 is attached to the second body part 5 by means of a strap 9. The shape of the straps 7, 9 and of the first and second position and orientation indicators 2, 4 is selected such that rotation about the local z-axis (e.g. the ^usz-axis or the ^LSz-axis) of the position and orientation indicators 2, 4 is prevented.

Fig. 4a schematically shows that, in case of a suitable design of the straps and of the position and orientation indicators, rotation about the local z-axis of the respective position and orientation indicators is prevented. However, Figs. 4b and 4c schematically show that the orientation of the position and orientation indicators 2, 4 with respect to their local y-axis (Fig. 4b) and with respect to their local x-axis (Fig. 4c) is not fixed, i.e. the respective position and orientation indicator is not aligned with respect to these axes. In other words, since the way of attachment restricts a rotation of the position and orientation indicator about its local z-axis, this alignment parameter can be assumed to be zero or some other constant. With respect to the local x-axis of the respective position and orientation indicator, in case of the body part being a (lower or upper) leg, rotation about the x-axis is not restricted (i.e. whether the position and orientation indicator is worn on the "front" of the leg or on the "side" of the leg). Similarly, with respect to the local y-axis of the position and orientation indicator, the local y-axis is not fixed, since it cannot be predetermined how cone shaped the patient's leg is. As a consequence of this possible misalignment of the respective position and orientation indicator with respect to the corresponding body part, two alignment parameters (the position of the x-axis and the position of the y-axis of the local position and orientation indicator coordinate system with respect to the local coordinate system of the corresponding body part) remain unknown.

Based on the assumption that rotation around the local z-axis is prevented, it can be started with calculating the difference between the instantaneous orientation of the joint ^wQs and its low-pass version |_^βs_| :

^SQ_VSΓ^WQS ®

\ (3) It should be noted that equation (3) expresses the difference in the position and orientation indicator coordinate system.

Since it has been assumed that the position and orientation indicator is attached to the body part such that the x-axis of the position and orientation indicator is along the direction (of extension) of the body part, it is also known that the twist rotation of the joint is along the x-axis of the quaternion

be decomposed into its three Euler angles heading, pitch, roll (in that order!) where the roll angle is the twist angle of the leg.

The roll angle φ (which is equal to the twist angle) can be calculated by:

where w, x, y and z are the w, x, y, and z components of S Q / s J •

Consequently, the quaternion ^LQrolls representing this roll angle is given by:

The quaternion ^LQrolls is the rotation from the S-coordinate system (with roll) to the L-coordinate system (without roll).

The rotation axis of the joint in the world coordinate system, excluding twist, can now easily be calculated with a principal component analysis of a sequence of:

Thus, it has been described above (equation (4)) how the instantaneous twist of the joint (with respect to the average twist) can be calculated. For physiotherapy exercises where it is important to either twist the joint or actually not twist the joint, this parameter can be used to give feedback to the patient. For example, for most hip exercises it is important that the twist of the leg stays below a certain threshold. The device according to the example is structured such that, if the angle φ exceeds a predefined threshold set in the control unit 100, a feedback is provided to the user. For example, this feedback can be provided as a message (e.g. visually on a suitable display or as speech via speakers) to the user like "Do not twist your leg". Alternatively, this feedback can also be signaled in a different way such as e.g. visually via suitable lamps or the like or as a specific sound. For example, the feedback can also be provided such that it has increasing intensity with increasing value of the angle φ.

Further, the device is structured to determine (as a second parameter) the direction in which a joint is rotated, excluding twist (using a principal component analysis of a sequence of equation (6)). In this way, it can e.g. be detected whether a leg is lifted to the front or to the side (in case of a hip joint movement). The control unit is further provided with suitable predefined information to determine whether the detected movement is desired or not. This can e.g. be realized by suitable threshold values set in a memory. When an undesired movement is detected, the device gives a feedback to the user, e.g. by providing a comment such as "Try to lift you leg sidewards, not forwards." or by another way of appropriate signaling.

A third parameter that can be extracted from a sequence of equation (6) is how well the "cloud" of data points (the points in Fig. 5) forms a straight line. If the points form a perfect line through the origin, it means that the patient is moving such that the joint has a constant rotation axis which is not varying over time. Effectively this means that the body part is moving within a plane (this plane does not need to be horizontal or vertical) which is desirable for most physiotherapy exercises. However, when the "cloud" of data points is broad, or does not form a straight line, the movement of the body part is not within a plane. The device according to the embodiment is structured such that the control unit 100 is adapted to evaluate this parameter. This can e.g. be achieved by monitoring the "cloud" of data points and calculating a standard deviation. If the standard deviation exceeds a threshold value set in the control unit 100, this triggers a feedback to the user like "Try to move your leg in a straight line" (for hip exercises). However, as mentioned with respect to the other parameters, the feedback can also be provided visually or as a sound or speech in a different way.

Further, for most physiotherapy exercises it can be judged that if the rotation axis has a comparably large vertical component the patient is also doing a wrong movement. In the case the position and orientation indicator measures its orientation with respect to the world (such as an electronic compass), a second position and orientation indicator may be worn on the body of the patient (for example on the chest) to determine in which compass direction the patient is standing, and thus relate the rotation axis of the joint not to a world fixed coordinate frame, but to a patient-fixed coordinate frame.

Although it has been described above with respect to the embodiment that only the position and orientation indicator readings of one position and orientation indicator (e.g. of either the first position and orientation indicator 2 or the second position and orientation indicator 4) are used for determining the rotation axis, by analyzing the movements of the position and orientation indicator with respect to a "world- fixed" reference system, the invention is not restricted to this. It should be noted that the description above implicitly assumed that the position and orientation of the joint of interest remain (substantially) constant over time. This assumption is reasonable e.g. in case of many physiotherapy exercises for the knee joint or the shoulder joint. However, at least in some cases this assumption is not valid. This is for instance the case for analyzing the rotation of the elbow joint while the shoulder joint is simultaneously moved or for analyzing the rotation of the knee joint while the hip joint is simultaneously moved. In these cases, the rotation axis of the joint of interest can still be determined using the method as described above, however with a slight adaptation which will be described in the following.

In the situation of the position and orientation of the joint of interest not remaining stationary during exercising, the rotation axis can still be determined if position and orientation indicators are worn on both sides of the joint during exercising (such as the first position and orientation indicator 2 worn "above" the knee joint and the second position and orientation indicator 4 worn "below" the knee joint for analyzing the rotation axis of the knee joint). However, in this case all the rotations of one of the position and orientation indicators have to be expressed with respect to the other one of the position and orientation indicators instead of with respect to a "world- fixed" reference system. As a consequence, the equations given above can still be used in this case; however the W signifying the "world- fixed" reference system has to be replaced by the corresponding information relating to the respective other position and orientation indicator. In other words, e.g. for analyzing the rotation of the knee joint, the position and orientation indicator readings of the first position and orientation indicator 2 have to be set in relation to the position and orientation indicator readings of the second position and orientation indicator 4 instead of in relation to the world- fixed system or vice versa.

Thus, according to the embodiment the rotation axis of a joint is determined from the position and orientation indicator readings of a position and orientation indicator without requiring alignment of the position and orientation indicator to the body part adjacent to the joint. From the thus-determined rotation axis, three different parameters are calculated. The rotation axis is expressed in a local coordinate frame for the twist, and in a world- fixed or global coordinate frame for the other two degrees of freedom. Further, these parameters are used to provide certain feedback to a patient or to a physiotherapist. The device for monitoring the movement is provided with a control unit adapted to perform the described steps.

Claims

CLAIMS:

1. Method for determining the rotation axis (R) of a joint, the method comprising the steps: attaching a position and orientation indicator (2, 4) to a body part (3, 5) adjacent to a joint; moving the joint and detecting resulting motions of the position and orientation indicator (2, 4) with respect to a reference system; determining instantaneous orientations (^WQs) of the position and orientation indicator (2, 4) at different points in time from the detected motions of the position and orientation indicator; determining an average orientation (|_^β_s J) of the position and orientation indicator from the detected motions of the position and orientation indicator; and determining a rotation axis (R) of the joint based on the instantaneous orientations and the average orientation.

2. Method according to claim 1, wherein the method comprises the step of calculating the differences (^SQ\_S \ ■ '_> ^WQ_\ w _\) between the instantaneous orientations and the average orientation, and the rotation axis of the joint is determined based on these differences.

3. Method according to any one of claims 1 or 2, wherein the reference system is a global fixed system.

4. Method according to any one of claims 1 or 2, wherein the reference system is a second position and orientation indicator (4, 2) attached to a body part on the opposite side of the joint.

5. Method according to any one of claims 1 to 4, wherein the position and orientation indicator (2, 4) is attached to the body part (3, 5) such that rotation about a first axis (^usz; ^LSz) of a local coordinate frame of the position and orientation indicator with respect to the body part is prevented.

6. Method according to claim 5, wherein the component (^LQrolls) of the rotation axis (R) along the first axis (^usx; ^LSx) of the local coordinate frame of the position and orientation indicator is determined from the differences between the instantaneous orientations of the position and orientation indicator and the average orientation of the position and orientation indicator.

7. Method according to claim 6, wherein the orientation of the rotation axis excluding the component of the rotation axis (R) along the first axis (^usx; ^LSx) of the local coordinate frame of the position and orientation indicator is determined.

8. Device for monitoring the movements of at least one body part comprising: at least one position and orientation indicator (2, 4) adapted to be attached to a body part (3, 5) which is adjacent to a joint; a control unit (100) adapted for detecting motions of the position and orientation indicator with respect to a reference system; wherein the control unit (100) is adapted to: determine instantaneous orientations (^WQs) of the position and orientation indicator at different points in time from the detected motions of the position and orientation indicator; determine an average orientation (|_^β_s J) of the position and orientation indicator from the detected motions of the position and orientation indicator; and determine a rotation axis (R) of the joint based on the instantaneous orientations and the average orientation.

9. Device according to claim 8, wherein the device comprises an attachment component (7, 9) for attaching the at least one position and orientation indicator (2, 4) to the body part (3, 5), the attachment component being structured such that rotation about a first axis (^usz; ^LSz) of a local coordinate frame of the position and orientation indicator with respect to the body part is prevented.

10. Device according to any one of claims 8 or 9, wherein the device comprises at least one further position and orientation indicator (4, 2) adapted to be attached to a further body part (5, 3) on the opposite side of the at least one joint.

11. Device according to any one of claims 8 to 10, wherein the device is a physiotherapy monitoring device.

12. Device according to any one of claims 8 to 11, wherein at least one position and orientation indicator is formed by a motion sensor.