WO2023033646A1

WO2023033646A1 - A method for automatically measuring a propulsive power applied to a pushrim of a wheelchair by a user of the wheelchair

Info

Publication number: WO2023033646A1
Application number: PCT/NL2022/050493
Authority: WO
Inventors: Maria Petronella VAN DIJK; Oscar BLIEK; Rienk Michiel Arjen VAN DER SLIKKE
Original assignee: Technische Universiteit Delft
Priority date: 2021-08-31
Filing date: 2022-08-30
Publication date: 2023-03-09
Also published as: NL2029083B1

Abstract

A computer-implemented method for measuring a propulsive power exerted on a pushrim of a wheelchair by a user of said wheelchair, comprising: determining a wheelchair energy state at a first moment, F1, and determining a wheelchair energy state at a second moment, F2, the energy states determined based upon a mass and a velocity and/or a mass and an acceleration of the wheelchair; estimating a mean resistance force acting on the wheelchair in between the first and second instance, Ff; determining a distance covered in between the first, F1, and the second, F2, instance, s automatically determining a duration of contact between a limb of the user and a pushrim of the wheelchair in between the first, F1 and the second, F2, time instance, t; and determining the propulsive power exerted on the pushrim of the wheelchair by said user based on F1, F2, Ff, s and t.

Description

Title: A method for automatically measuring a propulsive power applied to a pushrim of a wheelchair by a user of the wheelchair.

BACKGROUND

The present invention relates to a method for measuring a propulsive power exerted on a pushrim of a wheelchair by a user of said wheelchair, a method for the determination of a pushrim contact angle achieved by a user of a wheelchair, a detection device for determining contact between a limb of a wheelchair user and a pushrim of the wheelchair, and a wheelchair comprising such a detection device.

In the case of a wheelchair that is propelled by a user of the wheelchair himself/herself, the propulsion is effected by rotating the pushrims of the wheelchair that are mounted against the main wheels of the wheelchair in a forwards direction. Typically this is done with the hands of the user. As the pushrims are typically mounted to the main wheels at several points, it is not straightforward to measure the power exerted on the pushrims by the user while propelling the wheelchair. Typically, power exerted on the pushrims is defined by the force I torque applied on the pushrim, multiplied by the velocity I angular velocity, of the force application point. Whereas for e.g. bikes a simple power meter can be applied to a crank, this would not work for a wheelchair due to the multiplicity of contact points between the wheel and the pushrim.

At present, some measurements I estimates related to wheelchair propulsion by the user of the wheelchair can be made. However, all methods either have a very high margin of error, or make use of heavy and expensive devices that require a modification of the wheelchair.

For example, inertial measurement units applied to the wheelchair itself and/or applied to a chest of the person in the wheelchair and/or applied to upper arms of the person in the wheelchair are presently used to measure velocity and acceleration of the wheelchair or a respective body part of the user. When velocity and/or acceleration data is measured with an inertial measurement unit applied to the wheelchair only, useful data may be obtained regarding the net result of the wheelchair propulsion, but any measurements regarding the input of the user are impossible. It is noted that in the art of wheelchair measurements, the velocity and acceleration are typically expressed as a vector including a magnitude, i.e. an absolute value, and a direction. One of the main reasons that such measurements are impossible is that a wheelchair is not continuously propelled, but in a stepwise manner. The user grabs the wheels, rotates them by moving his or her hands, and then has to move the hands back to the initial position to be able to propel the wheelchair again. As such, there is a “propulsion phase” when the user contacts the pushrims, and a “recovery phase” when the user does not contact the pushrims. To make things more difficult from a measurement perspective, there may be propulsion I acceleration of the wheelchair in the recovery phase, when the user does not contact the pushrims, e.g. due to bodily movements of the user.

When an inertial measurement unit is additionally applied to the upper arms of the wheelchair user, velocity and/or acceleration data can be obtained on a “per stroke” basis, wherein a stroke is defined as a single push rotation effected by the user on the pushrim. However, more precise data cannot be obtained with any degree of confidence and data on a “per stroke basis” are not reliable and precise enough to determine peak loading of the user’s muscles.

Heavy and expensive devices, useable as specialized test equipment, are available to determine the propulsion power with more accuracy. However, such specialized test equipment is typically very expensive, very heavy, so that the results obtained are difficult to map to a “real-world” scenario without the test equipment present, and often requires a modification of the wheelchair as the pushrims have to be detached from the wheelchair. For experienced users, this may take 20 - 30 minutes. Additionally, such systems typically require calibration before they can be used, which again takes valuable time, and add a considerable mass (ca. 10 kg for the lightest-known system).

Injuries as a result of overloading of certain muscle groups are very frequent among wheelchair-dependent persons, especially athletes. In particular muscles around the shoulder joint are highly impacted when propelling a wheelchair with the hands of the wheelchair user. Such injuries are frequently seen among wheelchair athletes, but also among all other wheelchair-dependent persons. To prevent injuries from occurring seemingly randomly, it would be highly desirable if the power and/or force with which a wheelchair is propelled by a user can be monitored in real time. Additionally, it would be highly desirable if such measurements would be possible with equipment that can be applied to a wheelchair easily, that is cost-effective, and/or that has a low weight.

Other frequently-seen injuries relate to a suboptimal contact angle while propelling the wheelchair. Such injuries are seen in beginning wheelchair users and more experienced wheelchair users alike. Therefore, feedback about the contact angle during propulsion is crucial. However, at present such feedback is only possible in a simulated environment during a visitation of a physiotherapist, and not in a real-life context.

SUMMARY OF THE INVENTION

Accordingly, a first aspect of the present invention relates to a computer- implemented method for measuring a propulsive power exerted on a pushrim of a wheelchair by a user of said wheelchair, wherein the method comprises the steps of: determining a wheelchair energy state, e.g. a kinetic energy state, at a first instance in time, F1 , and determining a wheelchair energy state at a second instance in time, F2, the energy states being determined based upon a mass and a velocity and/or a mass and an acceleration of the wheelchair at said instance; estimating a resistance force acting on the wheelchair in between the first instance and the second instance, Ff; determining a distance covered in between the first, F1 and the second, F2, time instance, s automatically determining, a duration of contact between a limb of the user and a pushrim of the wheelchair, in between the first, F1 and the second, F2, time instance, t; and determining the propulsive power exerted on the pushrim of the wheelchair by said user based on F1 , F2, Ff, s and t.

Advantageously, with the method as presented herein an accurate determination of the propulsive force I power exerted by a user of the wheelchair can be determined, automatically and in real time. The invention is partly based on the insight that during propulsion of the wheelchair an acceleration may be obtained even after the user has stopped pushing. For example, when pushing the wheelchair the user typically moves his/her trunk forwards, to allow a larger push angle to be obtained and thus to propel the wheelchair more effectively. When accelerating the trunk back towards the back seat of the wheelchair, the wheelchair will in turn be propelled forwards. This acceleration may account for up to 35 % of the total propulsive force impulse when measuring on a “per stroke basis”, while it may take about 15 - 30 % of the total acceleration time for this latter acceleration effect as a result of trunk movement to be achieved. So, when calculating the force I power exerted by the user on the pushrim based on acceleration I velocity data of the wheelchair only, on the one hand the total amount of force applied with the hands of the user tends to be overestimated whereas the push duration tends to be underestimated. This leads to incorrect assumptions about wheelchair user capabilities, rolling resistance force and loading and this leads to an incorrect correlation of wheelchair-generated data and data generated outside of the wheelchair resulting in incorrect power calculations.

Accordingly, the present method, by taking into account the contact duration between the pushrim and the limb of the user, may greatly help therapists and coaches to accurately determine load capacity of wheelchair users and may prevent injuries from surfacing unexpectedly by allowing better tracking of said load capacity. By taking the contact duration into account automatically, the load capacity of wheelchair users can advantageously be tracked in real time, as they are in the wheelchair, to optimally prevent injuries. This brings monitoring on a daily basis into reach.

It is noted that previously it was not possible to determine the duration of limb contact time automatically. Accordingly, a further aspect of the present invention, to be discussed in more detail below, relates to how to determine said duration of contact automatically.

Another advantage of the present method is that more accurate tests can be performed when testing the effect of certain wheelchair components I variables on the performance of the wheelchair. When the input delivered by the user can be determined more reliably, this can be accounted for when interpreting the test results.

Advantageously, with the method as presented herein the propulsive power exerted by a user of the wheelchair can be measured in real time.

Advantageously, the method as presented herein will benefit both athletes and “regular” wheelchair users as more insight may be gained into the interaction between the user and the wheelchair. For athletes this may result in improving their performance, for “regular” users this may help when injuries are suffered from.

Advantageously, these measurements are possible with minimal inconvenience to the wheelchair user as he or she does not need to wear any special equipment.

The present invention relates to a method for measuring a propulsive power exerted on a pushrim of a wheelchair by a user of the wheelchair, the user exerting the force with his/her limb(s). It is noted that the word “limb” is not limited to hands with five fingers attached to it. As a wheelchair is virtually exclusively used by disabled persons, the wording “limb” includes a stump, as well as a hand with any number of fingers attached to it, e.g. between one and seven.

In accordance with the present invention, the time duration taken in between the first instance and the second instance may range anywhere from one or a few milliseconds to a typical stroke duration. In other words, the time duration in between the first and the second instance may be in between 1 millisecond to 3 seconds. This time duration will mainly result on the desired measurement accuracy.

It is noted that with measurement units applied to the wheelchair, the properties of the wheelchair-user system, e.g. velocity and acceleration, cannot be directly measured but instead the properties of the wheelchair are measured. However, it may be assumed that at least in the propulsion phase the velocity and acceleration of the wheelchair-user system is similar to the velocity and acceleration of the wheelchair.

In an embodiment of the present invention, an inertial measurement unit is used to determine the velocity and/or acceleration of the wheelchair at the first and/or the second instance. To obtain the most accurate results, the inertial measurement unit is preferably fixated to a main wheel axis of the wheelchair. In contrast, when an inertial measurement unit mounted to the user of the wheelchair would be used, the measurement results would be inaccurate to a smaller or larger degree as the user typically moves with respect to the frame of the wheelchair and this relative movement between user and wheelchair obstructs the accuracy of the measurements. As described in the below, advantageously an inertial measurement sensor worn by the user may be used in the context of the present invention in addition to a wheelchairmounted inertial measurement unit. In an embodiment of the present invention, the inertial measurement unit comprises an accelerometer, a gyroscope and/or a magnetometer. When all three of these components are present in the inertial measurement unit, as is the preferred embodiment, acceleration and velocity of the wheelchair can advantageously all be computed to allow the best possible and most accurate calculations regarding power produced by the user.

In an embodiment of the present invention, the resistance force acting on the wheelchair is determined while accounting for an orientation of the user in the wheelchair. It has been found by the present inventors that the rolling resistance quite significantly depends on the weight distribution of the wheelchair-user combination. When the user leans forward, a relatively high portion of the weight acts on the castor wheels of the wheelchair. The rolling resistance of these wheels differs from the rolling resistance of the main wheels. To account for these differences in resistance, for example tests can be performed with the user of the wheelchair in several, e.g. two, three, four or five different, positions to determine the rolling resistance of the wheelchair-user combination in these positions. For example, a curve fitting technique may be employed in between these data points to therefrom estimate with some accuracy the rolling resistance at any position of the user in the wheelchair.

Besides from the rolling resistance being non-constant, also the air resistance is non-constant. The air resistance however differs with speed mainly, and less with the position of the user (when typical wheelchair-accessible speeds are concerned). However, the speed can relatively easily be obtained with the inertial measurement unit. Typically, the rolling resistance is at least 2 - 3x higher than the air resistance and may be negligible at speeds below 20 km/h, such speeds being rarely reached with a wheelchair.

In yet alternative embodiments, the resistance force acting on the wheelchair may be estimated based on measurements made by force sensors attached to the wheelchair.

In an embodiment of the present invention, the orientation of the wheelchair user is determined via an inertial sensor worn by said user, preferably at the trunk of the user. The trunk position is a relatively accurate parameter to determine the weight distribution of the user in the wheelchair from. Conveniently, when using the presented method to determine the propulsive power applied by athletes, said athletes often wear a heart rate monitor, so that the inertial sensor can easily be attached to or integrated with a strap of said inertial sensor.

In an embodiment of the present invention, the duration of contact between the limb of the user and the pushrim of the wheelchair is measured by a capacitive sensor applied on the pushrim, the capacitive sensor extending across the entire outer circumference of the pushrim. For example, the capacitive sensor may be able to determine whether there is contact between the limb and the pushrim or whether there is no contact between the limb and the pushrim.

For example, the capacitive sensor may comprise two conductive elements that are spaced apart and arranged in parallel from each other, a voltage being applied on (only) one of the two conductive elements such that there is a voltage difference between the two conductive elements. When a limb of the user contacts the pushrim, the limb contacts both conductive elements and neutralizes the voltage difference between the two conductive elements by transferring the voltage from the one element to the other element through the limb, as the human skin is conductive for a voltage. When there is no voltage difference between the conductive elements the limb contacts the pushrim, when there is a voltage difference between the conductive elements there is no contact between the limb of the user and the pushrim.

By applying the conductive elements over the entire circumferential length of the pushrim, it can be determined with a high accuracy whether there is a contact between the limb and the pushrim.

For example, the capacitive sensor may comprise two or more, such as ten or more, fifteen or more or twenty or more conductive elements that are spaced apart from each other and arranged in series when seen in the longitudinal direction of the pushrim. Each of the conductive elements has a capacity sensor associated with it. By noting changes in the electrical field established by the capacity sensor, the capacitive sensor is able to detect whether a limb pushes on the pushrim (on the location of the sensor). The more sensors are used, the more accurate the position where a limb contacts the pushrim may be determined.

When using capacitive sensors working based on changes in the electrical field, it is required that the conductive elements are not contacted directly. Therefore, an isolating layer is arranged over the sensor. Compared to the above-mentioned example wherein the sensor is directly contacted this has the advantage that e.g. sweaty hands may not interfere with the measurements, and that in a real-life setting more accurate results can be obtained.

When using sensors on the pushrim that are arranged in series with respect to each other and that are spaced apart from each other in combination with an angular position sensor, it becomes more easy to determine the location on the pushrim where the limb initiates contact with the pushrim - and where the contact is terminated. As such, it becomes more easy to determine the push angle initiated by the user.

Tests performed by the inventors wherein the measurement results obtained with such a capacitive sensor were compared to video footage of a user pushing the wheelchair have indicated that the capacitive sensor is up to 95% accurate, or even more accurate. For a total push duration of between 0.8 and 1.4 seconds, the average difference between the two measurement methods was about 0.02 - 0.08 seconds, wherein the inventors noted that it was very difficult based on the available video footage to precisely establish when the limb contacted the pushrim and when the limb let go of the pushrim.

In an embodiment of the present invention, the capacitive sensor is integrated with a pushrim sleeve. This allows the capacitive sensor to be easily applied to and removed from a particular wheelchair, to allow its use on several wheelchairs in a highly effective manner.

In an embodiment of the present invention, the propulsive power exerted by the user is determined based on or with the use of the formula,

(F2-F1+(Ff*(s2-s1))) / t, provided that contact between the limb and the pushrim has been established in at least a part of the time period t, wherein

F2-F1 represents a change in kinetic energy state of the wheelchair-user system in between a first instance and a second instance;

Ff represents average the friction force on the wheelchair due to resistance forces acting thereon the wheelchair in between the first and the second instance; s2 - s1 represents the distance covered with the wheelchair in between the first instance and the second instance and t represents the duration of contact between a limb of the user and the pushrim in between the first instance and the second instance. Depending on the variables used to execute the calculation, one or more scaling or multiplicity factors will need to be applied in the above formula.

In the above equation it is assumed that in the time period in between the first and the second instance there is no propulsive force besides the input of the user.

Accordingly, in an alternative embodiment of the present invention the propulsive power exerted by the user is determined while compensating for any propulsive force generated by a backwards trunk movement of the wheelchair user. This may e.g. be accounted for by correlating an acceleration of the wheelchair as measured by the inertial measurement unit to the information obtained by the limb contact measurement. When there is an acceleration but no limb contact, the acceleration may be assumed to be resulting from a trunk movement, and such accelerations must be adequately accounted for I filtered to obtain accurate result about the input delivered with the hands I limbs of the user.

In an embodiment of the present invention, the method additionally comprises the step of automatically determining a pushrim contact angle achieved by the user of the wheelchair, said angle being determined based on the average wheelchair velocity during limb contact between the limb of the user and the pushrim of the wheelchair and the duration of limb contact. Advantageously, this may allow to track the progress and results in terms of technique used by wheelchair users, when it comes to the variable pushrim contact angle per stroke. This applies to beginners and advanced users alike, and can be done remotely and in real time. With remote it is meant that a doctor I physician I therapist I trainer does not need to be around the wheelchair user when tracking his/her progress, so that doctors’ appointments are needed less frequently.

This concept of remotely tracking the pushrim contact angle achieved by a user of the wheelchair solves the above-discussed technical problem also when the estimations regarding energy state and resistance force are not taken into account.

Accordingly, a second aspect of the present invention relates to a computer- implemented method for the automatic determination of a pushrim contact angle achieved by a user of a wheelchair, wherein the method comprises the steps of: determining a momentaneous velocity of the wheelchair; continuously detecting a presence of limb contact between a limb of the user and the pushrim of the wheelchair; determining a push duration based on a time difference between a start of the limb contact and an end of the limb contact; synchronizing wheelchair velocity and limb contact detection; and determining the pushrim contact angle based on the average wheelchair velocity and the push duration.

A person skilled in the art will understand that features and embodiments described as advantageous for the first aspect of the present invention may also be advantageously applied to the second aspect of the invention.

To allow the detection of contact between a limb of a wheelchair user and pushrim of a wheelchair, which previously could not be done reliably and easily, a third aspect of the present invention relates to a detection device for detecting contact between a limb of a wheelchair user and a pushrim of a wheelchair, the device to be used in combination with a wheelchair, the device comprising:

- an isolating layer configured to be arranged against an upper/outer surface of the pushrim;

- a pair of conductive elements, applied on top of the isolating layer at a distance from each other and in use shielded from the pushrim of the wheelchair by the isolating layer, the conductive elements extending over the entire length of the isolation layer;

- a battery configured to apply a voltage on one of the conductive elements; wherein the isolating layer is formed as or integrated with a pushrim sleeve, the pushrim sleeve having a circumferential length that substantially matches a circumferential length of the pushrim of the wheelchair.

As described in the above, the conductive elements may be able to determine whether there is contact between the limb and the pushrim or whether there is no contact between the limb and the pushrim although the conductive elements themselves may not be able to determine on their own a force I power that is applied to the pushrim. However, using the method in accordance with the first aspect of the invention may allow to determine this force / power. According to a first exemplary embodiment of the third aspect a voltage is applied on (only) one of the two conductive elements such that there is a voltage difference between the two conductive elements. When a limb of the user contacts the pushrim, the limb contacts both conductive elements and neutralizes the voltage difference between the two conductive elements by transferring the voltage from the one element to the other element through the limb, as the human skin is a conductor in and of itself. When there is no voltage difference between the conductive elements the limb contacts the pushrim; when there is a voltage difference between the conductive elements there is no contact between the limb of the user and the pushrim. By applying the conductive elements over the entire circumferential length of the pushrim, it can be determined with a high accuracy whether there is a contact between the limb and the pushrim.

Of course, the voltage applied to the conductive elements should not put the user of the wheelchair at risk. Therefore a voltage of at most 12 V is preferably used. More preferably this voltage may be lower than 5 V, such as about 3 or 4 V.

It should be noted that each time the conductive elements are contacted by the user, the voltage on the conductive element is transferred through the skin of the user. The above-mentioned voltages are far from dangerous to the user when applied repeatedly while, as is preferred, additionally not being noticed by the user.

The wheelchair pushrim is often made of a conductive material as well. To optimally measure when there is contact and when not, the conductive elements are isolated with respect to the pushrim by the isolation layer.

According to a second exemplary embodiment of the third aspect the capacitive sensor may comprise two or more, such as ten or more, fifteen or more or twenty or more conductive elements that are spaced apart from each other and arranged in series when seen in the longitudinal direction of the pushrim. Each of the conductive elements has a capacity sensor associated with it. By noting changes in the electrical field established by the capacity sensor, the capacitive sensor is able to detect whether a limb pushes on the pushrim (on the location of the sensor). The more sensors are used, the more accurate the position where a limb contacts the pushrim may be determined.

When using capacitive sensors working based on changes in the electrical field, it is required that the conductive elements are not contacted directly. Therefore, an isolating layer is arranged over the sensor. Compared to the above-mentioned first example wherein the sensor is directly contacted this has the advantage that e.g. sweaty hands may not interfere with the measurements, and that in a real-life setting more accurate results can be obtained.

When using sensors on the pushrim that are arranged in series with respect to each other and that are spaced apart from each other in combination with an angular position sensor, in addition to determining whether contact is made, it becomes more easy to determine the location on the pushrim where the limb initiates contact with the pushrim - and where the contact is terminated. As such, it becomes more easy to determine the push angle initiated by the user.

To allow easy installation and de-installation of the detection device on a wheelchair, the isolation layer and conductive elements are formed as or integrated with a pushrim sleeve. Such sleeves are already well known to wheelchair users and can be very light-weight. Both these factors will likely help in the practical acceptability and implementation of the present invention in the wheelchair using community.

Tests performed by the inventors wherein the measurement results obtained with such a detection device were compared to video footage of a user pushing the wheelchair have indicated that the detection device is up to 95% accurate or even more accurate. For a total push duration of between 0.8 and 1.4 seconds, the average difference between the two measurement methods was about 0.02 - 0.08 seconds, wherein the inventors noted that it was very difficult, based on the video footage, to precisely establish when the limb contacted the pushrim and when the limb let go of the pushrim..

For example, the distance between the conductive elements of the pair of conductive elements may be between 2 mm and 4 cm.

For example, the conductive elements may comprise copper material, and in particular may be made of a copper-nickel alloy. Alternatively, the conductive elements may be made of any other equivalent material that is both conductive and robust against tearing.

A fourth aspect of the present invention relates to a wheelchair comprising the detection device as described in the above. These and other aspects and embodiments of the present invention will now be elucidated further with reference to the below figures. In the figures, the same reference numerals are used to denominate same or like parts and components.

BRIEF DESCRIPTION OF THE FIGURES

Figures 1A - 1C schematically illustrate a user in a wheelchair, the user being in different positions with respect to the wheelchair;

Figure 2A schematically illustrates a first embodiment of a detection device in accordance with the present invention;

Figure 2b schematically illustrates a second embodiment of a detection device in accordance with the present invention; and

Figure 3 schematically illustrates a force diagram as established with the use of the present method and detection device.

DETAILED DESCRIPTION OF THE FIGURES

With reference to Figures 1A, 1 B and 1C initially, which are discussed together, a user U is shown in a wheelchair 100. The wheelchair comprises castor wheels 31 at the front thereof, main wheels 32 at the rear thereof, and a pushrim 33 at both sides connected to the main wheels 32 to allow a user U to manipulate and rotate the main wheels 32 to propel the wheelchair 100 forwards and/or backwards. The wheelchair 100 further comprises a frame 34 to which an inertial measurement unit 21 is mounted, and a seatback 35. It is noted that in Figures 1A - 1C the unit 21 is shown as attached to the frame, for clarity purposes mainly. In real life, the unit 21 is more likely to be attached to a rotation axis of the main wheels 32.

The inertial measurement unit 21 preferably comprises a gyroscope, an accelerometer and a magnetometer (not shown). Using the inertial measurement unit 21 the velocity and acceleration of the wheelchair 100 can be obtained. Another inertial sensor 11 is connected to a trunk of the user U, i.e. is worn by the user U. The inertial sensor 11 worn by the user preferably also comprises an accelerometer, gyroscope and magnetometer. As will be appreciated from a comparison of the figures, when the user U moves backwards and forwards in the seat 35 of the wheelchair 100, the relative positions of the inertial measurement unit 11 attached to the frame 34 of the wheelchair 100 changes with respect to the position of the inertial sensor 11 worn by the user II. From this change in relative position of the inertial measurement unit 21 and inertial sensor 11 the relative orientation of the user II in the seat 35 may be obtained I estimated. This is important to estimate the resistance of the user-wheelchair combination, as will be explained in more detail in the below.

In the shown examples, a forward propulsion by the user is illustrated. To that end, the user II grasps the pushrims 33 with his/her limbs, typically hands, and moves the pushrims 33 forwards. To allow the maximum rotation of the pushrims 33 to be achieved, the user II may rotate his/her body forwards, so that an approximately 90-degree contact angle may be obtained.

In Figure 1A the state S1 is shown wherein the user II rests against the backseat 35 and just grabs the pushrims 33 to initiate a rotation of the pushrims 33. In this initial state S1 , the wheelchair will have a velocity, which might be zero, positive or may even be negative. Likewise, the user-wheelchair combination has an acceleration in the initial state S1. These parameters may be obtained by the inertial measurement unit 11 , e.g. continuously. Relatively constant will be the weight of the user-wheelchair combination. From these parameters, an energy state of the wheelchair-user combination at that particular instance in time may be obtained.

In Figure 1 B the state S2 is shown wherein the user II is halfway his/her push rotation. The user II has moved forwards with respect to the backseat 35 of the wheelchair and the inertial sensor 11 worn by the user II, with moves along with the user relative to the wheelchair 100, has moved towards the inertial measurement unit 21 attached to frame 34 of the wheelchair 100, which remains at the same position relative to the wheelchair 100. In the second state, possibly corresponding to a second instance in time, the velocity and acceleration of the wheelchair can again be obtained by inertial measurement unit 21 attached to the wheelchair frame 34. Again, from these parameters, possibly in combination with the mass of the user-wheelchair combination that typically is not significantly changed in between states S1 and S2, an energy state of the user-wheelchair combination may be obtained in this second state S2.

Likewise, a resistance or friction force will act on the wheelchair-user combination in between the first state S1 and the second state S2. This friction force may be estimated when the weight of the user-wheelchair combination is known, when the underground is known, when the pressure in the main wheels 32 and the type of castor wheels 32 is known, and when the position of the user II in the wheelchair 100 is known or estimated such that the weight distribution of the wheelchair 100 is known and when the velocity of the wheelchair 100 is known.

Based on the difference in energy state in the initial state S1 and the latter state S2, in combination with the friction force acting on the wheelchair-user combination, it can be determined how much energy (force, power) has been applied to the system to bring it from state S1 to state S2. However, when generalizing to any state S1 and any state S2, not known per se is how this energy is added, especially when the method is carried out remotely and the user II cannot be seen.

The inventors have realized that, in principle, the energy can be added in a number of ways.

Firstly, the energy might be added by a motor. However, the present invention is not related to motor-driven wheelchairs so that is ruled out.

Secondly, the energy might be added by a person pushing the wheelchair. However, the present invention is not related to wheelchair propelled by a person other than the user itself, so that option is ruled out as well.

Thirdly, the energy state might be lower in state S2 compared to state S1. In that case, the propulsive force applied is lower than the friction force, which makes it likely that no energy is added by the user.

Fourthly, energy may be added by manually pushing the pushrims 33. When pushing the pushrims 33, a contact between the hands of the user II and the pushrims 33 may be detected, and a contact duration may be determined. From the contact duration and the amount of energy added, the propulsive power and/or force follows.

Fifthly, when the user II pushes back his/her trunk towards the seatback 35 from the situation in Figure 1 B, the wheelchair 100 will also be propelled forwards as the action of the user II moving backwards will result in a reaction of the wheelchair moving forwards. In this case, energy is added to the system without the hands of the user II contacting the pushrims 33, so that this energy cannot directly be related to arms I shoulder muscles power of the user II exerting force I power on the pushrims 33, but instead to trunk muscles of the user II exerting force I power on the wheelchair 100 as a whole. In previous methods, the fourth and fifth option of adding energy to the wheelchair-user system tended to be combined. However, this led to incorrect results regarding power and force being added by the user II via the pushrims 33. Now that it is possible to distinguish between energy added by pushing the pushrims 33 and energy added by movement of the trunk, through the determination of contact between limbs of the user II and the pushrim 33, it becomes possible to determine more accurately than before how much power and force is added to the wheelchairuser combination in which stage of propulsion. For example, when contact between the hands and the pushrim has been established for at least a portion of the time in between the first instance and the second instance, the propulsive power exerted by a user on the pushrims of a wheelchair 100 may be determined based upon:

(energy state at instance two - energy state at instance one + (average friction force in between instances one and two * distance covered in that time)) I duration of contact between hands and pushrim

To optimally estimate the friction force acting on the wheelchair-user combination, it may e.g. be assumed that while the user II is actively propelling the wheelchair 100 by rotating the pushrims 33 with his/her hands, the weight distribution of the wheelchair and user is similar to the weight distribution when the user is sitting with his/her back against the backseat 35, as while the user II is contacting and propelling the pushrims 33 the majority of his/her weight will act on the main wheels 32. As soon as the user II lets go of the pushrim 33, the weight distribution of the user-wheelchair combination changes. In particular, when the user II releases the pushrims 33 from a position as indicated in Figures 1 B and 1 C, a relatively large portion of the weight rests on the castor wheels 31. As the rolling resistance of the castor wheels 31 is different from the rolling resistance of the main wheels 32, this weight distribution has an effect on the rolling resistance of wheelchair 100 as a whole. By tracking the position of the user II in the wheelchair 100 from comparing position data obtained by the inertial measurement sensor 11 worn by the user II to the position data obtained by the inertial measurement sensor 21 attached to the wheelchair, the position of the user II in the wheelchair 100 may be estimated. From initial tests covering a few positions, the rolling resistance at any position may be estimated, e.g. by fitting a curve of any number of polynomials through the data points. Finally shown in Figure 1C is a third state, in which the user II moves forward maximally to exert the maximum amount of force on the wheelchair pushrims 33 per stroke. The calculation as described in the above may be made at any instance, e.g. once or multiple times in between state S1 and S2, once or multiple times in between states S2 and S3, or once in between state S1 and S3, or once in between state S1 and the next state S1 following upon state S3 as indicated, when the user II has moved back into state S1 again from state S3.

Another type of information that is of interest when measuring wheelchair propulsion efficiency and/or wheelchair user loading, is the contact angle of the pushrim obtained by the user. Now that it is possible to obtain a contact duration between a hand of the user and a pushrim 33 of the wheelchair 100, as well as a velocity of the wheelchair 100 / an average velocity of the wheelchair 100 in between a start of the pushing rotation and an end of the pushing rotation, said contact angle can quite easily be obtained from the average velocity of the wheelchair 100 and the contact duration between the hand of the user II and the pushrim 33 of the wheelchair 100.

One possible embodiment of a detection device 1 that may be used to measure contact between the hand of the user and the pushrim is shown in Figure 2A. Shown in Figure 2A is an isolation layer 2, which may form the basic layer of the detection device. In particular, the isolation layer 2 will in use be applied on the pushrim of the wheelchair, such that the pushrim, which is typically made of a conductive material, will not interfere with the measurements. Applied on the isolation layer 2 are two lines 3, 4 made of a conductive material. These lines will in use be visible and form the top of the device, so that a user, when pushing the pushrim, contacts both lines. The lines will, in use, be applied on the entire outer circumference of the pushrim, so that the position where the user pushes the pushrim does not influence the measurement results. Applied on one conductive line, in this case the line indicated with reference numeral 4, a positive voltage of e.g. less than 12 V is applied via battery 5 and wire 9C. The second line of conductive material, indicated with reference numeral 3, is connected to the ground via wires 9D and 9B so that no voltage is applied on that particular element. The two conductive elements 3, 4 are separated from each other on the isolation layer 2 so that any voltage applied on the first conductive element 4 cannot be transferred to the second conductive element 3. This changes when a hand of a wheelchair user contacts the two elements 3, 4, as the wheelchair user will have to do in order to be able to propel the wheelchair 100. As the human skin is a conductor of itself, the hand will make sure that the voltage applied on the first conductive element 4 is transferred to the second conductive element 3. Transistor 7 ensures that the voltage on line 9D is grounded instead of shorted and measurement line 9A will notice a voltage on line 9E, the measurement line indicating that the hand of the user contacts the pushrim. Chip 6 determines from these measurement a total contact duration.

An alternative embodiment of a detection device 1 that may be used to measure contact between the hand of the user and the pushrim is shown in Figure 2B. Shown in Figure 2B is a first isolation layer 2A, applied on the pushrim of the wheelchair. This ensures that the pushrim, which is typically made of a conductive material, will not interfere with the measurements. Applied on the first isolation layer 2A are a plurality of capacitive sensors 3, 4. As is assumed known to those skilled in the relevant art, a capacitive sensor 3, 4 may output a signal when an object - metallic or non-metallic - is brought in its vicinity. In order for this to work, a voltage must be applied on the sensor 3, 4. Preferably said voltage is below 12 V. In this case, when a limb of a user will grab the pushrim 33 (including the detection device 1) the capacitive sensor 3, 4 outputs a signal. Hence, based on the length of the signal the contact duration between the limb and the pushrim 33 can be determined. Further, when used in combination with an angular motion sensor, it can be determined at what position the pushrim 33 is grabbed and when it is released, so that the push angle may be determined reliably. As may be derived from Figure 2B the sensors 3, 4 may be arranged along the entire circumferential length of the pushrim 33, the sensors 3, 4 individually being spaced apart along the longitudinal direction of the pushrim 33 so that it can be determined accurately which of the sensors 3, 4 is grabbed when pushing the pushrim 33. Provided on top of the capacitive sensors 3, 4 is a second layer of isolating material 2B. As such, the capacitive sensors 3, 4 will in use be invisible. The second layer of isolating material 2B guarantees the working of the capacitive sensor 3, 4 when relying on changes in the magnetic field associated with the sensor 3, 4. For the embodiment of both figures 2A and 2B it holds that as conductive material of the conductive elements 3, 4 e.g. a copper material can be used, in particular a copper-nickel alloy or any other conductive material.

The physical wires of the detection device 1 can e.g. be mounted to a spoke of the main wheel 32, the chip and other control elements of the detection device 1 being mounted to the frame 34 of the wheelchair or near the axle of the main wheels at a stationary, i.e. non-rotating position.

Advantageously the conductive device is formed as or integrated with a pushrim sleeve to be applied on the pushrim. Pushrim sleeves are already familiar to wheelchair users such that the detection device will have a familiar feel and implementation and acceptability is maximized. Further, the pushrim sleeve hardly adds any weight to the wheelchair and is easily installed on a wheelchair without major modifications.

Figure 3 shows some results that are obtained while using the abovedescribed method and the above-described detection device. The results are obtained automatically and could be obtained in real time, of course allowing for some processing time of in the order of seconds or less (depending on how efficient the coding is, how soon the data is transported from the sensors to the processor, etc.).

The figure shows, one propulsive cycle of the user on the wheelchair, over time. The propulsive stroke starts when the hands of the user grab the pushrims, continues as the user rotates the pushrims, further continues as the user releases the pushrims, further continues as the user moves his/her trunk back to the initial position and stops when the user again grabs the pushrims to start the next propulsion cycle.

Shown in grey is the contact between the pushrims and the limbs of the user. This contact time is present for about 40% of the propulsion cycle. It is noted that this percentage is not always 40%, but may depend on the intensity with which the user is propelling the wheelchair.

Shown in the solid line is the propulsion force acting on the wheelchair. It is well noticeable in the figure that the propulsion force continues after the user has let go of the pushrims with his/her hands. Shown in the dashed line is the trunk angle of the user with respect to the vertical. Visible from the figure is that the rotation of the trunk from the (near) horizontal position back to the (near) vertical position results in a propulsive force being applied on the wheelchair without a contact between the hands and the pushrims being established.

Knowing the data as derivable from Figure 3, in combination with the other data as described in this application, allow for accurate and, possibly, real-time calculations regarding propulsive forces and power applied on the wheelchair by the user thereof.

Claims

1. A computer-implemented method for measuring a propulsive power exerted on a pushrim of a wheelchair by a user of said wheelchair, wherein the method comprises the steps of: determining a wheelchair energy state at a first instance in time, F1 , and determining a wheelchair energy state at a second instance in time, F2, the energy states being determined based upon a mass and a velocity and/or a mass and an acceleration of the wheelchair at said instance; estimating a mean resistance force, Ff, acting on the wheelchair in between the first instance and the second instance; determining a distance, s, covered in between the first, F1 , and the second, F2, time instance, automatically determining a duration of contact, t, between a limb of the user and a pushrim of the wheelchair in between the first, F1 and the second, F2, time instance, using a capacitive sensor applied on the pushrim, the capacitive sensor extending across the entire outer circumference of the pushrim; and determining the propulsive power exerted on the pushrim of the wheelchair by said user based on F1 , F2, Ff, s and t.

2. The computer-implemented method according to claim 1 , wherein an inertial measurement unit is used to determine the velocity and/or acceleration of the wheelchair at the first, F1 , and/or the second, F2, instance.

3. The computer-implemented method according to claim 2, wherein the inertial measurement unit comprises an accelerometer, a gyroscope, and/or a magnetometer.

4. The computer-implemented method according to any one of the preceding claims, wherein the resistance force acting on the wheelchair, Ff, is determined while accounting for an orientation of the user in the wheelchair.

5. The computer-implemented method according to claim 4, wherein the orientation of the user is determined via an inertial sensor worn by the user, preferably at his/her trunk.

6. The computer-implemented method according to any one of the preceding claims, wherein the capacitive sensor is integrated with a pushrim sleeve.

7. The computer-implemented method according to any one of the preceding claims, wherein the propulsive power exerted by the user is determined based on the formula

(F2-F1+(Ff*(s2-s1))) / t, provided that contact between the limb and the pushrim had been established in the time period t.

8. The computer-implemented method according to any one of the preceding claims, wherein the propulsive power exerted by the user is determined while compensating for any propulsive force generated by a backwards trunk movement of the wheelchair user.

9. The computer-implemented method according to any one of the preceding claims, additionally comprising the step of automatically determining a pushrim contact angle achieved by the user of the wheelchair, said angle being determined based on the average wheelchair velocity during limb contact between the limb of the user and the pushrim of the wheelchair and the duration of limb contact.

10. A computer-implemented method for the automatic determination of a pushrim contact angle achieved by a user of a wheelchair, wherein the method comprises the steps of: determining a momentaneous velocity of the wheelchair; continuously detecting a presence of limb contact between a limb of the user and the pushrim of the wheelchair using a capacitive sensor applied on the pushrim, the capacitive sensor extending across the entire outer circumference of the pushrim; determining a push duration based on a time difference between a start of the limb contact and an end of the limb contact; synchronizing wheelchair velocity and limb contact detection; and determining the pushrim contact angle based on the average wheelchair velocity and the push duration.

11. A detection device for detecting contact between a limb of a wheelchair user and a pushrim of a wheelchair, the device to be used in combination with a wheelchair, the device comprising:

12. The detection device according to claim 11 , wherein the distance between the conductive elements of the pair of conductive elements is between 2 mm and 4 cm.

13. The detection device according to claim 11 or 12, wherein the conductive elements are spaced apart in a longitudinal direction of the detection device.

14. The detection device according to any one of the claims 11 - 13, wherein the conductive elements comprise copper, and in particular are made of a copper-nickel alloy.

15. A wheelchair comprising the detection device according to any one of the claims 11 - 14.