Device for detecting mechanical force at multiple regions of a surface

Abstract

The present disclosure relates to a device for detecting mechanical force at multiple regions of a surface, the device comprising: a plurality of sensor bodies, wherein each of the plurality of sensor bodies comprises: a deformable material; and a piezoelectric sensor configured to detect a displacement of a region of the surface upon deformation of the deformable material by the region of the surface; wherein the piezoelectric sensor of a first one of the plurality of sensor bodies is isolated from the piezoelectric sensor of a second one of the plurality of sensor bodies such that deformation of the piezoelectric sensor of the first one of the plurality of sensor bodies does not cause deformation of the piezoelectric sensor of the second one of the plurality of sensor bodies.

Description

[0001] DEVICE FOR DETECTING MECHANICAL FORCE AT MULTIPLE REGIONS OF A SURFACE

[0002] FIELD

[0003] The present disclosure relates to a device for detecting mechanical force at multiple regions of a surface, and in particular for detecting displacement of multiple regions beneath a subject’s skin, so that a movement made by the subject or a physiological state of the subject can be recognised.

[0004] BACKGROUND

[0005] Current devices and systems for detecting a subject’s gestures for virtual reality (VR) applications include VR gloves, camera tracking, and surface electromyography (EMG). VR gloves use accelerometers to track a subject’s movement and detect their gestures. VR gloves also implement haptic feedback motors in order to provide haptic feedback for subjects. However, VR gloves are an expensive solution as a result of the complex hardware that they require. VR gloves can also be uncomfortable over time, as they are heavy, clunky and can quickly become sweaty. Camera tracking is another method used in VR applications for detection of gestures. However, this method suffers from blindspots, when a tracked object (e.g. a subject’s hand) is obscured (e.g. by the subject themselves). Surface EMG involves measuring electrical activity of muscles. However, human subjects typically have low muscle mass at the wrist, meaning that surface EMG is ineffective for a wrist-based solution for gesture detection.

[0006] Existing devices that use sensors to detect movement of tendons in order to identify gestures typically implement off-the-shelf components. These devices typically receive complex signals, meaning that complex filtering algorithms are required in order to process the sensor outputs and identify a subject’s gestures.

[0007] Accordingly, there exists a need for an improved device for detecting displacement of one or more tendons beneath a subject’s skin, so that a gesture made by the subject can be recognised. More generally, there exists a need for an improved device for detecting displacement at or beneath a subject’s skin (or for detecting displacement of any other surface). SUMMARY

[0008] This summary introduces concepts that are described in more detail in the detailed description. It should not be used to identify essential features of the claimed subject matter, nor to limit the scope of the claimed subject matter.

[0009] According to one aspect of the present disclosure, there is provided a device for detecting mechanical force at multiple regions of a surface, the device comprising: a plurality of sensor bodies, wherein each of the plurality of sensor bodies comprises: a deformable material; and a piezoelectric sensor configured to detect a displacement of a region of the surface upon deformation of the deformable material by the region of the surface; wherein the piezoelectric sensor of a first one of the plurality of sensor bodies is isolated from the piezoelectric sensor of a second one of the plurality of sensor bodies such that deformation of the piezoelectric sensor of the first one of the plurality of sensor bodies does not cause deformation of the piezoelectric sensor of the second one of the plurality of sensor bodies.

[0010] Isolating the piezoelectric sensors from one another greatly simplifies the process of classifying the outputs from the piezoelectric sensors, because displacement of one piezoelectric sensor does not result in any interference being detected at the other piezoelectric sensors. Therefore, the signal from each piezoelectric sensor reflects the true displacement of a region of the surface beneath the piezoelectric sensor, rather than some or all of the signal reflecting displacement of adjacent piezoelectric sensors. Simplifying the classification process means that gestures can be detected more easily, and with reduced processing requirements.

[0011] The piezoelectric sensor of a first one of the plurality of sensor bodies may be isolated from the piezoelectric sensor of each of the other ones of the plurality of sensor bodies such that deformation of the piezoelectric sensor of the first one of the plurality of sensor bodies does not cause deformation of the piezoelectric sensor of each of the other ones of the plurality of sensor bodies.

[0012] The device may further comprise a rigid backing. The rigid backing prevents the force applied to one of the sensor bodies from being transferred to any of the other sensor bodies. The deformable material may be connected to the rigid backing. The plurality of sensor bodies may be attached to the rigid backing such that adjacent ones of the plurality of sensor bodies are separated by a gap. The gap allows the piezoelectric sensors to be displaced independently of one another. The gap may be less than 1 mm, or more preferably less than 0.4 mm. Alternatively, the plurality of sensor bodies may be attached to the rigid backing such that adjacent ones of the plurality of sensor bodies are configured to slide against each other. The slidable movement of adjacent piezoelectric sensors allows the piezoelectric sensors to be displaced independently of one another.

[0013] The device may comprise a plurality of rigid backings. Adjacent ones of the plurality of rigid backings may be attached together. Each of the plurality of rigid backings may be attached to one or more respective sensor bodies of the plurality of sensor bodies. Each of the plurality of rigid backings may be attached to two or more respective sensor bodies of the plurality of sensor bodies. Providing a plurality of rigid backings attached together allows the sensor bodies of the device to conform to a curved surface such as a subject’s wrist.

[0014] The deformable material may be an elastomeric material. The deformable material may be a first deformable material. Each of the plurality of sensor bodies may further comprise a second deformable material. The use of two deformable materials allows the piezoelectric sensor to be displaced relative to the rigid backing when a force is applied to the first deformable material. The piezoelectric sensor may be sandwiched between the first deformable material and the second deformable material.

[0015] Sandwiching the piezoelectric sensor between the deformable materials allows the piezoelectric sensor to be implemented parallel to the rigid backing, which minimises the size of the sensor body. Alternatively, the piezoelectric sensor may be embedded in the deformable material.

[0016] The plurality of sensor bodies may comprise an array of sensor bodies attached to the rigid backing. The array of sensor bodies may comprise at least two rows of sensor bodies. Implementing multiple rows of the array also allows the absolute position of the array on a subject’s wrist to be determined. Sensor bodies in a first row of the array may be offset from sensor bodies in a second row of the array. By offsetting one row of sensor bodies from another row of sensor bodies, the likelihood of a sensor body of the array being positioned directly over a region of a surface that is displaced (e.g. a region of a subject’s skin that is displaced by a tendon) is increased. The diameter of each of the plurality of sensor bodies may be less than or equal to 5 mm, less than or equal to 4 mm, less than or equal to 3 mm, or less than or equal to 2 mm. Minimising the sensor body diameter maximises the density of sensor bodies that may be implemented on the device.

[0017] The device may be configured to conform the plurality of sensor bodies to a curved surface. Accordingly, the device can be used to detect mechanical force at regions of a subject’s body such as the wrist.

[0018] The device may further comprise a casing configured to constrain lateral movement of each of the plurality of sensor bodies. Each of the plurality of sensor bodies may protrude from a respective one of a plurality of openings in the casing. Constraining lateral movement of the sensor bodies minimises the shear force applied to the piezoelectric sensors, which prevents damage to the piezoelectric sensors and therefore provides increased robustness of the sensor bodies, while also simplifying the classification of the outputs from the piezoelectric sensors.

[0019] Each of the plurality of sensor bodies may comprise a rigid button. The rigid button may be attached to the deformable material. The rigid button of each of the plurality or sensor bodies may protrude from a respective one of a plurality of openings in the casing. The rigid button allows lateral movement of the sensor body to be constrained.

[0020] The piezoelectric sensor may comprise a piezoelectric film. A piezoelectric film is preferred because it deflects more in response to an applied force than, for example, a solid body, and therefore provides a stronger signal in response to an applied force.

[0021] The piezoelectric sensor may be a first piezoelectric sensor. Each of the plurality of sensor bodies may comprise a second piezoelectric sensor. Implementing two piezoelectric sensors in each sensor body allows more information on surface displacement to be detected by the piezoelectric sensors.

[0022] The device may further comprise a processor configured to process an output from the piezoelectric sensor of each of the plurality of sensor bodies. Processing the outputs from the piezoelectric sensors using a processor of the device reduces latency compared with transmitting the sensor measurements to a remote device for processing. The processor may be configured to implement a trained classifier configured to classify one or more movements based on the output from the piezoelectric sensor of each of the plurality of sensor bodies.

[0023] BRIEF DESCRIPTION OF FIGURES

[0024] Specific embodiments are described below by way of example only and with reference to the accompanying drawings, in which:

[0025] FIG. 1 shows a schematic diagram of a device comprising a plurality of sensor bodies according to a first example.

[0026] FIG. 2 shows a schematic exploded view of a sensor body of the device shown in FIG. 1.

[0027] FIG. 3 shows a schematic diagram of an alternative sensor body to the sensor body shown in FIG. 2.

[0028] FIG. 4 shows a schematic diagram of a further alternative sensor body to the sensor body shown in FIG. 2.

[0029] FIG. 5 shows a schematic diagram of a plurality of sensor bodies of a device according to a second example.

[0030] FIG. 6 shows a schematic diagram of an enclosure of the device according to the second example.

[0031] FIG. 7 shows a schematic diagram of a curved enclosure assembly of the device according to the second example.

[0032] FIG. 8 shows a schematic diagram of the device according to the second example.

[0033] FIG. 9 shows a schematic diagram of a plurality of sensor bodies of a device according to a third example.

[0034] FIG. 10 shows a schematic diagram of the device according to the third example. FIG. 11A shows a distribution of normalised energy values detected by piezoelectric sensors in response to a subject raising their thumb with the palm of their hand flat on a surface.

[0035] FIG. 11B shows a distribution of normalised energy values detected by piezoelectric sensors in response to a subject raising their index finger with the palm of their hand flat on a surface.

[0036] FIG. 11C shows a distribution of normalised energy values detected by piezoelectric sensors in response to a subject raising their middle finger with the palm of their hand flat on a surface.

[0037] FIG. 11 D shows a distribution of normalised energy values detected by piezoelectric sensors in response to a subject raising their ring finger with the palm of their hand flat on a surface.

[0038] FIG. 11 E shows a distribution of normalised energy values detected by piezoelectric sensors in response to a subject raising their little finger with the palm of their hand flat on a surface.

[0039] DETAILED DESCRIPTION

[0040] Implementations of the present disclosure are explained below with particular reference to detection of displacement of tendons beneath the skin of a subject’s wrist, in order to classify one or more gestures made by the subject. It will be appreciated, however, that the implementations described herein are also applicable to detection of displacement of other regions of a subject’s skin (e.g. temples, ankle) in order to classify movements of certain body parts of the subject. More generally, it will be appreciated that the implementations described herein are not limited to detecting displacement of human skin, and are capable of detecting displacement of multiple regions of any solid surface.

[0041] FIG. 1 is a schematic diagram of a device 100 comprising a plurality of sensor bodies 110. As shown schematically in FIG. 1 , the plurality of sensor bodies 110 are attached to a rigid backing 102 of the device, such that each sensor body 110 is separated from an adjacent sensor body 110 by a gap 104. Given that tendons at the wrist are typically 2-4 mm in diameter, it is preferable to have a separation of at least one order of magnitude lower than the diameter of a tendon. For example, a preferred range for the gap 104 between adjacent sensor bodies 110 may be greater than 0 mm and less than or equal to 1 mm for a device 100 intended for use at a subject’s wrist, and more preferably greater than 0 mm and less than or equal to 0.4 mm. The gap 104 may be larger for devices 100 used at other locations on a subject’s skin, such as at the ankle or near the elbow. FIG. 1 shows that the plurality of sensor bodies 110 can be provided in the form of a sensor array.

[0042] As shown in FIG. 1 and the exploded schematic diagram of FIG. 2, each sensor body 110 comprises a first deformable material 112, a piezoelectric sensor 114, and a second deformable material 116. In the example shown in FIGS. 1 and 2, each of the first deformable material 112 and the second deformable material 116 has a disc shape (meaning a cylindrical shape with a small axial dimension), while the piezoelectric sensor 114 is provided in the form of a circular film. For example, the first and second deformable materials 112, 116 may each be provided in the form of discs (cylinders) having a diameter of 5 mm and a height of 1 mm, whereas the piezoelectric sensor 114 is provided in the form of a circular film having a diameter of 4 mm and a negligible thickness (e.g. 28 pm).

[0043] The piezoelectric sensor 114 has a smaller diameter than both the first deformable material 112 and the second deformable material 116. The piezoelectric sensor 114 is sandwiched between the first deformable material 112 and the second deformable material 116. In the example shown in FIGS. 1 and 2, the piezoelectric sensor 114 is sandwiched between the first deformable material 112 and the second deformable material 116 so that no part of the piezoelectric sensor 114 protrudes from the side of the sensor body 110 (for example, the piezoelectric sensor 114 can be centred with respect to the first deformable material 112 and the second deformable material 116 such that the centre of the piezoelectric sensor 114 is aligned with the centres of the first and second deformable materials 112, 116). Keeping the piezoelectric sensor 114 within the footprint of the deformable materials 112, 116 helps protect the piezoelectric sensor 114 from being disconnected from its connecting wires. These wires are placed on the extremities of the piezoelectric sensor 114, and would potentially be exposed if the piezoelectric sensor 114 was not disposed within the footprint of the deformable materials 112, 114. Sandwiching the piezoelectric sensor 114 between the deformable materials 112, 116 means that the piezoelectric sensor 114 is parallel to the rigid backing 102, which minimises the size of the sensor body 110 (meaning that a more compact device 100 can be provided). Layers of an elastomeric adhesive such as silicone sealant (not shown) may be provided between the first deformable material 112 and the piezoelectric sensor 114, and between the piezoelectric sensor 114 and the second deformable material 116. The second deformable material 116 is attached to the rigid backing 102.

[0044] The use of two deformable materials 112, 116 allows the piezoelectric sensor 114 to be displaced relative to the rigid backing 102 when a force is applied to the first deformable material 112 (e.g. by a tendon or muscle beneath a subject’s skin). The displacement of the piezoelectric sensor 114 generates a voltage that can be detected using circuitry (not shown in FIGS. 1 and 2). Although not shown in FIG. 2, an electrical connection is provided between a circuit board and either side of the piezoelectric sensor 114. For example, a ground connection may be provided between the circuit board and the side of the piezoelectric sensor 114 adjacent to the first deformable material 112, while a positive connection may be provided between the circuit board and the side of the piezoelectric sensor 114 adjacent to the second deformable material 116. The detected voltages can be processed in order to identify which piezoelectric sensors 114 of the sensor bodies 110 are displaced by a particular displacement beneath a surface.

[0045] The first and second deformable materials 112, 116 may be resiliently deformable materials such as elastomeric materials. In one example, the first deformable material 112 is the same as the second deformable material 116. An example of a suitable elastomeric material is room-temperature-vulcanising (RTV) silicone with a shore hardness of 15A. Other suitable elastomeric materials include rubbers, neoprene, and soft thermoplastic elastomers. As shown in FIG. 2, the piezoelectric sensor 114 is preferably provided in the form of a piezoelectric film. A piezoelectric film is preferred because it deflects more in response to an applied force than, for example, a solid body, and therefore provides a stronger signal in response to an applied force. An example of a suitable piezoelectric film is polyvinylidene fluoride (PVDF) film.

[0046] Referring back to FIG. 1 , it can be seen that the gap 104 between a first sensor body 110a and a second sensor body 110b allows a piezoelectric sensor 114a of the first sensor body 110a to be displaced without causing displacement of a piezoelectric sensor 114b of the second sensor body 110b. In other words, each of the piezoelectric sensors 114 is isolated from the other piezoelectric sensors 114. This isolation is provided by (i) the gap between the sensor bodies 110, which allows the piezoelectric sensors 114 to be displaced independently of one another; (ii) the rigid backing 102, which prevents the force applied to one of the sensor bodies 110 from being transferred to any of the other sensor bodies 110. Isolating the piezoelectric sensors 114 from one another greatly simplifies the process of classifying the outputs from the piezoelectric sensors 114, because displacement of one piezoelectric sensor 114 does not result in any interference being detected at the other piezoelectric sensors 114. Therefore, the signal from each piezoelectric sensor 114 in the array of sensor bodies 110 reflects the true displacement of a region of the surface beneath the piezoelectric sensor 114, rather than some or all of the signal reflecting displacement of adjacent piezoelectric sensors 114. Simplifying the classification process means that gestures can be detected more easily, and with reduced processing requirements. The signal processing can be carried out using analogue or digital circuits, a microcontroller or other chips such as FPGAs, or on a computer (e.g. using GPUs, CPUs, etc.). Where gesture classification is carried out at the device 100 (e.g. by a processor of the device), reducing the processing requirements means that the amount of processing carried out at the device 100 is reduced, thereby prolonging battery life.

[0047] FIG. 3 is a schematic diagram of an alternative sensor body 140 that may be used in the device 100. For example, one or more of the sensor bodies 110 of the device 100 may be replaced with a sensor body 140 having the construction shown in FIG. 3. The sensor body 140 comprises a first deformable material 142, a first piezoelectric sensor 144, a second deformable material 146, a second piezoelectric sensor 148, and a third deformable material 150.

[0048] Each of the deformable materials 142, 146, 150 of the sensor body 140 has the same construction as the deformable materials 112, 116 of the sensor body 110, while each of the piezoelectric sensors 144, 148 of the sensor body 140 has the same construction as the piezoelectric material 114 of the sensor body 110. Therefore, the first piezoelectric sensor 144 is sandwiched between the first deformable material 142 and the second deformable material 146, while the second piezoelectric sensor 148 is sandwiched between the second deformable material 146 and the third deformable material 150.

[0049] Each of the deformable materials 142, 146, 150 may be the same material (e.g. silicone with a shore hardness of 15A). Alternatively, one of the deformable materials 142, 146, 150 may be a different material to one or more of the other deformable materials 142, 146, 150. For example, the first deformable material 142 may have a lower hardness than the second deformable material 146, and the second deformable material 146 may have a lower hardness than the third deformable material 150. In this way, the first piezoelectric sensor 144 may be sensitive to small changes in displacement of the surface, while the second piezoelectric sensor 148 is sensitive to larger changes in displacement of the surface. This means that more information on surface displacement can be detected by the piezoelectric sensors 144, 148.

[0050] FIG. 4 is a schematic diagram of a further alternative sensor body 170 that may be used in the device 100. For example, one or more of the sensor bodies 110 of the device 100 may be replaced with a sensor body 170 having the construction shown in FIG. 4. The sensor body 170 comprises a single deformable material 172 and a piezoelectric sensor 174 embedded in the single deformable material 172. In the example shown in FIG. 4, the deformable material 172 has a cylindrical shape, having an axial dimension that exceeds that of the deformable materials 112, 116 of the sensor body 110. As with the sensor bodies 110, 140, the deformable material may be silicone with a shore hardness of 15A.

[0051] In contrast to the sensor bodies 110, 140, the piezoelectric sensor 174 of the sensor body 170 has a rectangular shape (i.e. is a rectangular film with negligible thickness), and is embedded in the deformable material 172 in a vertical orientation. In other words, the piezoelectric sensor 174 (which has a negligible axial dimension) is aligned with the longitudinal axis of the deformable material 172 (in contrast to the sensor bodies 110, 140, in which the piezoelectric sensors 114, 144, 148 are aligned with the radial axes of the deformable materials 112, 116, 142, 146, 150). Consequently, when a force is applied to a distal end of the deformable material 172, a bending motion is applied to the piezoelectric sensor 174. FIG. 4 also shows the electrical connections from either side of the piezoelectric sensor 114 to a circuit board (not shown).

[0052] FIG. 5 shows four of sensor bodies 210 coupled to a rigid backing 202 of a device 200, where the rigid backing 202 is provided in the form of a printed circuit board (PCB). Each of the sensor bodies 210 has the same construction as the sensor bodies 110 described above (i.e. includes the first deformable material 112, the piezoelectric sensor 114 and the second deformable material 116), except that each of the sensor bodies 210 also includes a rigid button 218 attached to the first deformable portion 112. The rigid button 218 has a cylindrical shape of similar diameter to the first deformable material 112. The rigid button 218 transfers an applied force (e.g. by displacement of the surface) to the first deformable material 112 and consequently to the piezoelectric sensor 114. The rigid button 218 allows lateral movement of the sensor body 210 to be constrained, when used in conjunction with a casing (shown in FIG. 6).

[0053] As described above, electrical connections are provided between the sides of the piezoelectric sensors 114 of the sensor bodies 210 and the PCB. As shown in FIG. 5, an array comprising two rows of two sensor bodies 210 is attached to the rigid backing 202. The sensor bodies 210 of one row are offset from the sensor bodies 210 of the other row. By offsetting one row of sensor bodies 210 from another row of sensor bodies 210, the likelihood of a sensor body 210 of the array being positioned directly over a region of a surface that is displaced (e.g. a region of a subject’s skin that is displaced by a tendon) is increased. As shown in FIG. 5, the rigid backing 202 has a shape equivalent to two conjoined offset equiangular quadrilateral shapes. This means that a ‘step’ is provided along opposite sides of the rigid backing (as best shown in FIG. 7).

[0054] FIG. 6 shows a casing 220 of the device 200, which is attached to the rigid backing 202. In the example shown in FIG. 6, the casing 220 includes two side portions 222, each of which is coupled to one end of the rigid backing 202 (i.e. a side of the rigid backing 202 without the ‘step’), and also to the face of the rigid backing 202 to which the sensor bodies 210 are attached. Each of the two side portions 222 is integral with a surface facing portion 224 of the casing 220. The surface facing portion 224 has the same shape as the rigid backing 202 (i.e. equivalent to two conjoined equiangular quadrilaterals), meaning that a ‘step’ is provided along the sides of the surface facing portion 224 that are not connected to the side portions 222. The two side portions 222 extend at right-angles to the surface facing portion 224, meaning that the casing 220 has a U-shaped cross-section. When the rigid backing 202 is attached to the casing 220, the rigid backing 202 is parallel to the surface facing portion 224. Accordingly, the rigid backing 202 and the casing 220 together form an enclosure 234 in which the sensor bodies 210 are located.

[0055] As shown in FIG. 6, the surface facing portion 224 of the casing 220 includes four openings 226. Each of the openings 226 is aligned with one of the sensor bodies 210. Each of the openings 226 has a diameter that is slightly greater than the diameter of the rigid buttons 218, thereby allowing the each of the rigid buttons 218 to pass through a corresponding opening 226. When the device 200 is assembled, each of the rigid buttons 218 protrudes from a corresponding one of the openings 226. The rigid buttons 218 therefore protrude from the enclosure 234 defined by the casing 220 and the rigid backing 202. For example, the rigid buttons 218 may protrude by 2 mm from the enclosure 234. The openings 226 constrain the rigid buttons 218 so that they can only move substantially normal to the rigid backing 202 (where “substantially normal” is used to recognise that the slight difference in diameter between the rigid buttons 218 and the openings 226 may result in the rigid buttons 218 moving towards or away from the rigid backing 202 at a small angle to the normal). Constraining the rigid buttons 218 in this way minimises lateral movement of the rigid buttons 218, thereby minimising the shear force applied to the piezoelectric sensors 114. Minimising the shear force applied to the piezoelectric sensors 114 prevents damage to the piezoelectric sensors 114 (e.g. the potential for piezoelectric sensors 114 to become disconnected from their connecting wires). Accordingly, minimising the shear force applied to the piezoelectric sensors 114 provides increased robustness of the sensor bodies 210. In addition, shear forces may complicate the data collected from the piezoelectric sensors 114. Accordingly, minimising the shear force applied to the piezoelectric sensors 114 simplifies the classification of gestures (e.g. carried out by a processor of the device 200).

[0056] FIG. 6 also shows that two further holes 228 extend through the casing 220. The holes 228 extend through the casing 220 in a direction parallel to the ends of the rigid backing 202. The holes 228 allow multiple enclosures 234 to be attached together to form a curved enclosure assembly 230 (as shown in FIG. 7), for example using wires that extend through the holes 228 of each of a plurality of casings 220. Accordingly, each enclosure 234 forms a ‘link’ of the curved enclosure assembly 230. In addition, the curved enclosure assembly 230 provides attachment of the plurality of rigid backings 202 together (i.e. via the attachment of the corresponding casings 220). A stiff non-stretchable rubber-coated wire can be used to attach the casings 220 together using the holes 228 extending through the casings 220. The casings 220 can be attached together as tightly as possible, leaving negligible free movement for the ‘links’ of the curved enclosure assembly 230. By using a non-stretchable wire, the curved enclosure assembly 230 can conform to a curved surface such as a subject’s wrist, but translation of each ‘link’ (casing 220) away from the surface is prevented. Therefore, the curved enclosure assembly 230 provides rigidity that prevents the deformation of a sensor body 210 on one link from causing deformation of an adjacent link, thereby providing isolation of piezoelectric sensors 114 on adjacent links from one another.

[0057] The curved enclosure assembly 230 provides conformity with the curved surface of a subject’s wrist, thereby allowing the sensor bodies 210 to be implemented in the device 200, which in this example is provided in the form of a wrist strap (as shown in FIG. 8). Providing a curved enclosure assembly 230 with links allows the device 200 to be used with a range of different wrist sizes and shapes. FIG. 8 shows that the device 200 can be formed by attaching a flexible strap 232 to either end of the curved enclosure assembly 230.

[0058] Also shown in FIG. 8 are the plurality of rigid buttons 218 extending through the corresponding openings 226 in the casings 220. FIG. 8 further shows the sensor bodies 210 of one row of the array of sensor bodies 210 being offset from the sensor bodies 210 of an adjacent row of the array of sensor bodies 210. In the example shown in FIG. 8, eight enclosures 234 are attached together, in order to provide a sixteen-by-two array of sensor bodies 210. The shape of the surface facing portion 224 of the casing 220 (specifically, the sides having the ‘step’) maintains the offset between sensor bodies 210 in adjacent rows when the enclosures 234 are attached together, while minimising a gap 204 between adjacent sensor bodies 210 when the enclosures 234 are attached together.

[0059] FIG. 9 shows a casing assembly of an alternative device 300 in which a plurality of alternative sensor bodies 310 is implemented. As with the devices 100, 200 described above, the device 300 includes a rigid backing 302 in the form of a PCB. The device further comprises a casing 320 that houses the plurality of sensor bodies 310. Each of the plurality of sensor bodies 310 comprises a first deformable material 312, a piezoelectric sensor 314, and a second deformable material 316. In this example, the first deformable material 312 has a diameter of 3.75 mm and a height of 1 mm, while the second deformable material 316 has a diameter of 3.75 mm.

[0060] Each sensor body 310 has a similar construction to the sensor bodies 110 shown in FIGS. 1 and 2, except that the second deformable portion 316 of some (or all) of the sensor bodies 310 has an angled face to which the piezoelectric sensor 314 is attached. In addition, as shown in FIG. 9, the height (i.e. axial dimension) of the second deformable portion 316 of some sensor bodies 310 differs from the height of the second deformable portion 316 of other sensor bodies 310. In particular, the height of the second deformable portion 316 of the sensor bodies 310 at the ends of the array of sensor bodies 310 is greater than the height of the second deformable portion 316 of the sensor bodies 310 at the middle of the array of sensor bodies 310. More specifically, as shown in FIG. 9, the height of the second deformable portion 316 of the sensor bodies 310 gradually reduces from one edge of the array to the middle of the array, and then gradually increases from the middle of the array to the other edge of the array. FIG. 9 shows that these changes in height of the second deformable portion 316 result in the first deformable portions 312 of the sensor bodies 310 providing a substantially curved surface that confirms to the surface being displaced (e.g. a subject’s wrist). In this context, “substantially” curved refers to the overall shape of the first deformable portions 312, while recognising that the first deformable portions 312 have flat surfaces and that there is a gap 304 between first deformable portions 312 of adjacent sensor bodies 310. As an alternative, a curved rigid backing could be used in conjunction with sensor bodies with the same height, to provide the substantially curved surface.

[0061] As shown in FIG. 9, the casing 320 comprises a plurality of cylindrical openings 326 that fully house the second deformable portions 316 and the piezoelectric sensors 314 of the sensor bodies 310. In addition, the walls of the cylindrical openings 326 comprise an electrical connection (e.g. a ground connection) to the piezoelectric sensor 314. The casing 320 also partially houses the first deformable portions 312, as shown by the slight overlap between the casing 320 and the first deformable portions 312.

[0062] This means that the casing 320 constrains the first deformable portions 312 so that they can only move substantially normal to the rigid backing 302, thereby minimising lateral movement of the first deformable portions 312 and consequently minimising the application of shear forces to the piezoelectric sensors 314.

[0063] As shown in FIG. 10, the casing assembly shown in FIG. 9 may be housed within a moulding 334 that encloses the casing assembly. A flexible strap may then be attached to the moulding 334 so that the device 300 can be worn on a subject’s wrist. FIG. 10 also clearly shows that the array of sensor bodies 310 includes four rows of sensor bodies 310, in which the sensor bodies 310 of one row are offset from the sensor bodies 310 of an adjacent row.

[0064] FIG. 11A shows a distribution of normalised energy values detected by piezoelectric sensors of an eight-by-two array of sensor bodies (e.g. piezoelectric sensors 314 of the array of sensor bodies 310 shown in FIGS. 9 and 10) in response to a test subject raising their thumb with the palm of their hand flat on a surface. FIGS. 11 B to 11 E show distributions of normalised energy values detected by the piezoelectric sensors in response to the test subject raising, respectively, their index finger, middle finger, ring finger and little finger. As shown in FIGS. 11 A to 11 E, raising each digit results in distinct energy distributions. The distinctness of the energy distributions means that the action of raising each digit can be clearly distinguished from the actions of raising other digits, based on the energy distribution detected by the piezoelectric sensors 104.

[0065] The distinctness of the energy distributions resulting from raising different digits also allows a machine-learning classifier to be trained to detect different gestures based on the energy distribution detected by the piezoelectric sensors. For example, a machinelearning classifier can be trained and tested using ten raises of each digit (i.e. a total of fifty samples), where 90% of the samples are used for training the classifier, and 10% are used for testing, with five-fold cross-validation. Training a support vector machine in this way resulted in a classifier with 100% test accuracy. The machine-learning classifier may be implemented by a processor of the device that detects the forces at the surface, or by a different device (e.g. a remote device such as a PC, smartphone, or other computing device). To reduce latency, the classifier may be implemented by a processor of the device that detects the forces. The outputs of the classification algorithm may then be transmitted to a remote device (e.g. via Bluetooth or another wireless connection).

[0066] Variations and modifications to the systems and methods described herein are set out in the following paragraphs.

[0067] It will be appreciated that various features of the examples described above may be implemented in other examples. For example, the device 300 may comprise a casing with openings through which the sensor bodies protrude in the same way as the device 200. In this case, the sensor bodies of the device 300 may comprise rigid buttons that protrude through the openings. As another example, multiple instances of the rigid backings of the device 100 may be attached together to conform the sensor bodies of the device 100 to a curved surface such as a subject’s wrist.

[0068] The devices described herein may additionally comprise haptic feedback motors in order to provide haptic feedback to subjects (e.g. in response to detection of particular gestures). In addition, infrared LEDs may be used for tracking the location of the device.

[0069] Although the sensor bodies described above have specific diameters (e.g. 5 mm, 3.75 mm), it will be appreciated that the diameter of the sensor bodies is not limited to the specific values given above. In general, smaller diameter sensors are preferable for maximising resolution, but have less favourable signal-to-noise ratios. In particular, the sensor bodies may have a diameter that is less than or equal to 5 mm, less than or equal to 4 mm, less than or equal to 3 mm, or less than or equal to 2 mm, in order to maximise the density of sensor bodies that may be implemented in the array. The minimum diameter of the sensor bodies is limited by the connection of the PVDF film to the wire used for connection to the circuit board. Copper tape and / or conductive adhesives may be used for connecting the PVDF film to the wire.

[0070] Moreover, although circular piezoelectric sensors and disc shaped deformable materials are used in the implementations described above, other shapes of the sensors and deformable materials may be implemented. For example, piezoelectric sensors and deformable materials with rectangular, square, triangular, hexagonal, or any other shape may be implemented.

[0071] As an alternative to implementing a gap between sensor bodies, the sensor bodies may be in contact with one another, but arranged to slide against each other, such that there is low friction between adjacent sensor bodies. A sensor body that slides against an adjacent sensor body can be deformed independently of the adjacent sensor body.

[0072] As an alternative to implementing piezoelectric sensors that are parallel to the rigid backing (e.g. as shown in FIGS. 1-3 and 5-8) or perpendicular to the rigid backing (e.g. as shown in FIG. 4), the piezoelectric sensor may be provided in the form of a curved piezoelectric film that is embedded in one or more deformable materials. Curving the piezoelectric film allows non-normal forces (i.e. shear forces) to be captured by the sensor body. Classification may also be used in order to distinguish between various gestures. In addition, although separate deformable materials are shown in the examples described with reference to FIGS. 1 to 3, it will be appreciated that the piezoelectric sensors of these examples may alternatively be embedded in a single deformable material.

[0073] Although various examples described above utilise a PCB as the rigid backing, it will be appreciated that the sensor bodies may be attached to a different rigid backing (e.g. defined by the casing). In this case, the piezoelectric sensors may be electrically connected to a PCB that does not function as the rigid backing.

[0074] The electrical connections to the piezoelectric sensor may be provided in a number of different ways. For example, wires may extend through one or more of the deformable materials in order to provide a connection to the piezoelectric sensor. Alternatively, a coating may be applied to the casing and / or the rigid backing, in order to provide part (or all) of the electrical connection.

[0075] While specific array sizes are described above, it will be appreciated that any array size from a two-by-one array upwards may be implemented in order to detect displacement of distinct regions of a surface. Increasing the number of columns of the array increases the number of regions of the surface for which the displacement can be detected. Increasing the number of rows of the surface also increases the detection capability. As explained above, offsetting adjacent rows increases the likelihood of a sensor body directly overlying the region of the surface that is being displaced. The likelihood of a sensor body directly overlying the region of the surface that is being displaced can also be increased using other array arrangements such as a random distribution or circular distribution.

[0076] Implementing multiple rows of the array also allows the absolute position of the array on a subject’s wrist to be determined. This is because tendons have a specific orientation, meaning that an orientation of the device with respect to the subject’s wrist can be determined based on the signals detected at each row of the array.

[0077] In addition to classifying gestures based on the energy distributions detected by the piezoelectric sensors, other features of the signal detected by the piezoelectric sensors may be used to classify gestures. For example, the shape, rise time, fall time, rise amplitude and / or fall time of the signal may be used to distinguish between different gestures.

[0078] Although the devices described above are used for detecting displacement of tendons beneath the skin of a subject’s wrist, the displacement of other regions of a subject’s skin may also be measured using the devices described above. For example, the devices described above may be mounted on a subject’s head in order to detect displacement of the skin at the subject’s temples. Based on the signal responses detected by placing a device at a subject’s temples, different eye movements may be detected. For example, the device may be capable of determining a direction in which the subject is looking. In general, the devices described above are not limited to detection of displacement of a surface resulting from tendon activation, and may be used to detect displacement of a surface resulting from muscle activation and blood flow (e.g. to detect a pulse). In such examples, movements other than gestures may be detected. For example, eye movements may be detected by a device worn on the subject’s head (e.g. around the ears), or foot movements may be detected by a device worn around the subject’s ankle. A pulse may also be measured at a suitable location on the skin (e.g. the wrist), in order to calculate a heart rate of the subject.

[0079] In addition, the use of the devices described above is not limited to detecting displacement of human skin. More generally, the devices described above are capable of detecting multiple regions of a surface.

[0080] The described methods may be implemented using computer executable instructions. A computer program product or computer readable medium may comprise or store the computer executable instructions. The computer program product or computer readable medium may comprise a hard disk drive, a flash memory, a read-only memory (ROM), a CD, a DVD, a cache, a random-access memory (RAM) and / or any other storage media in which information is stored for any duration (e.g., for extended time periods, permanently, brief instances, for temporarily buffering, and / or for caching of the information). A computer program may comprise the computer executable instructions. The computer readable medium may be a tangible or non-transitory computer readable medium. The term “computer readable” encompasses “machine readable”.

[0081] The singular terms “a” and “an” should not be taken to mean “one and only one”. Rather, they should be taken to mean “at least one” or “one or more” unless stated otherwise. The word “comprising” and its derivatives including “comprises” and “comprise” include each of the stated features, but does not exclude the inclusion of one or more further features.

[0082] The above implementations have been described by way of example only, and the described implementations are to be considered in all respects only as illustrative and not restrictive. It will be appreciated that variations of the described implementations may be made without departing from the scope of the invention. It will also be apparent that there are many variations that have not been described, but that fall within the scope of the appended claims.

Claims

CLAIMS:

1. A device for detecting mechanical force at multiple regions of a surface, the device comprising: a plurality of sensor bodies, wherein each of the plurality of sensor bodies comprises: a deformable material; and a piezoelectric sensor configured to detect a displacement of a region of the surface upon deformation of the deformable material by the region of the surface; wherein the piezoelectric sensor of a first one of the plurality of sensor bodies is isolated from the piezoelectric sensor of a second one of the plurality of sensor bodies such that deformation of the piezoelectric sensor of the first one of the plurality of sensor bodies does not cause deformation of the piezoelectric sensor of the second one of the plurality of sensor bodies.

2. A device according to claim 1 , wherein the device further comprises a rigid backing, and wherein the plurality of sensor bodies are attached to the rigid backing such that adjacent ones of the plurality of sensor bodies are separated by a gap.

3. A device according to claim 1 , wherein the device further comprises a rigid backing, and wherein the plurality of sensor bodies are attached to the rigid backing such that adjacent ones of the plurality of sensor bodies are configured to slide against each other.

4. A device according to claim 2 or claim 3, wherein the device comprises a plurality of rigid backings, wherein adjacent ones of the plurality of rigid backings are attached together, and wherein each of the plurality of rigid backings is attached to one or more respective sensor bodies of the plurality of sensor bodies.

5. A device according to any of claims 1 to 4, wherein the deformable material is an elastomeric material.

6. A device according to any of claims 1 to 5, wherein the deformable material is a first deformable material, and wherein each of the plurality of sensor bodies further comprises a second deformable material.

7. A device according to claim 6, wherein the piezoelectric sensor is sandwiched between the first deformable material and the second deformable material.

8. A device according to any of claims 1 to 6, wherein the piezoelectric sensor is embedded in the deformable material.

9. A device according to any of claims 2 to 8 when dependent on claim 2 or claim 3, wherein the plurality of sensor bodies comprises an array of sensor bodies attached to the rigid backing.

10. A device according to claim 9, wherein the array of sensor bodies comprises at least two rows of sensor bodies.

11. A device according to claim 10, wherein sensor bodies in a first row of the array are offset from sensor bodies in a second row of the array.

12. A device according to any of claims 1 to 11 , wherein the diameter of each of the plurality of sensor bodies is less than or equal to 5 mm.

13. A device according to any of claims 1 to 12, wherein the device is configured to conform the plurality of sensor bodies to a curved surface.

14. A device according to any of claims 1 to 13, further comprising a casing configured to constrain lateral movement of each of the plurality of sensor bodies.

15. A device according to claim 14, wherein each of the plurality of sensor bodies protrudes from a respective one of a plurality of openings in the casing.

16. A device according to any of claims 1 to 15, wherein the piezoelectric sensor comprises a piezoelectric film.

17. A device according to any of claims 1 to 16, the piezoelectric sensor is a first piezoelectric sensor, and wherein each of the plurality of sensor bodies comprises a second piezoelectric sensor.

18. A device according to any of claims 1 to 17, further comprising a processor configured to process an output from the piezoelectric sensor of each of the plurality of sensor bodies.

19. A device according to claim 18, wherein the processor is configured to implement a trained classifier configured to classify one or more movements based on the output from the piezoelectric sensor of each of the plurality of sensor bodies.

Images